GPU自动调度卷积层
GPU自动调度卷积层
本文对GPU使用自动调度程序。
与依靠手动模板定义搜索空间的基于模板的autotvm不同,自动调度程序不需要任何模板。用户只需要编写计算声明,无需任何调度命令或模板。自动调度程序可以自动生成一个较大的搜索空间,在该空间中找到良好的调度。
本文以卷积层为例。
本文无法在Windows或最新版本的macOS上运行。要使其运行,需要将本文的内容包装在一个if __name__ == "__main__":块中。
import os
import numpy as np
import tvm
from tvm import te, auto_scheduler, topi
from tvm.topi.testing import conv2d_nchw_python
定义计算
首先,定义卷积层的计算。该函数应返回输入/输出张量的列表。通过这些张量,自动调度器可以获得整个计算图。
@auto_scheduler.register_workload
def conv2d_layer(N, H, W, CO, CI, KH, KW, stride, padding):
data = te.placeholder((N, CI, H, W), name="data")
kernel = te.placeholder((CO, CI, KH, KW), name="kernel")
bias = te.placeholder((1, CO, 1, 1), name="bias")
conv = topi.nn.conv2d_nchw(data, kernel, stride, padding, dilation=1, out_dtype="float32")
out = topi.nn.relu(conv + bias)
return [data, kernel, bias, out]
创建搜索任务
然后,为resnet中的最后一个卷积层创建搜索任务。
target = tvm.target.Target("cuda")
# Use the last layer in ResNet-50
N, H, W, CO, CI, KH, KW, strides, padding = 1, 7, 7, 512, 512, 3, 3, (1, 1), (1, 1)
task = auto_scheduler.SearchTask(
func=conv2d_layer, args=(N, H, W, CO, CI, KH, KW, strides, padding), target=target
)
# Inspect the computational graph
print("Computational DAG:")
print(task.compute_dag)
输出:
Computational DAG:
data = PLACEHOLDER [1, 512, 7, 7]
pad_temp(i0, i1, i2, i3) = tir.if_then_else(((((i2 >= 1) && (i2 < 8)) && (i3 >= 1)) && (i3 < 8)), data[i0, i1, (i2 - 1), (i3 - 1)], 0f)
kernel = PLACEHOLDER [512, 512, 3, 3]
compute(nn, ff, yy, xx) += (pad_temp[nn, rc, (yy + ry), (xx + rx)]*kernel[ff, rc, ry, rx])
bias = PLACEHOLDER [1, 512, 1, 1]
T_add(ax0, ax1, ax2, ax3) = (compute[ax0, ax1, ax2, ax3] + bias[ax0, ax1, 0, 0])
compute(i0, i1, i2, i3) = max(T_add[i0, i1, i2, i3], 0f)
接下来,为自动调度程序设置参数。这些参数主要指定在搜索过程中如何进行测量。
- measure_ctx启动不同的测量过程以提供隔离。保护主进程免受测量期间GPU崩溃的影响,避免其它运行时冲突。
- min_repeat_ms定义每次测量中一次“重复”的最小持续时间。这样可以预热GPU,对于获得准确的测量结果是必不可少的。通常,建议值> = 300毫秒。
- num_measure_trials是在搜索过程中可以使用的测量试验的数量。为了快速演示,在本文中仅进行了10次试用。在实践中,1000是使搜索收敛的一个好值。可以根据自己的时间预算进行更多试验。
- 此外,还用RecordToFile将测量记录转储到文件conv2d.json中。测量记录可用于最好地查询历史记录,恢复搜索以及以后进行更多分析。
- 有关更多参数auto_scheduler.TuningOptions, 请参见auto_scheduler.LocalRPCMeasureContext。
log_file = "conv2d.json"
measure_ctx = auto_scheduler.LocalRPCMeasureContext(min_repeat_ms=300)
tune_option = auto_scheduler.TuningOptions(
num_measure_trials=10, # change this to 1000 to achieve the best performance
runner=measure_ctx.runner,
measure_callbacks=[auto_scheduler.RecordToFile(log_file)],
verbose=2,
)
输出:
Get devices for measurement successfully!
运行搜索
现在准备好所有输入。开始搜索,让自动调度程序发挥作用。经过一些测量试验之后,可以从日志文件中加载最佳调度并应用它。
# Run auto-tuning (search)
task.tune(tune_option)
# Apply the best schedule
sch, args = task.apply_best(log_file)
# Kill the measurement process
del measure_ctx
输出:
可以降低调度以在自动调度后查看IR。自动调度程序可以正确执行优化,包括多层平铺,协作提取,展开和算子融合。
print("Lowered TIR:")
print(tvm.lower(sch, args, simple_mode=True))
输出:
Lowered TIR:
primfn(data_1: handle, kernel_1: handle, bias_1: handle, compute_1: handle) -> ()
attr = {"global_symbol": "main", "tir.noalias": True}
buffers = {compute: Buffer(compute_2: Pointer(float32), float32, [1, 512, 7, 7], []),
kernel: Buffer(kernel_2: Pointer(float32), float32, [512, 512, 3, 3], []),
bias: Buffer(bias_2: Pointer(float32), float32, [1, 512, 1, 1], []),
data: Buffer(data_2: Pointer(float32), float32, [1, 512, 7, 7], [])}
buffer_map = {data_1: data, kernel_1: kernel, bias_1: bias, compute_1: compute} {
attr [IterVar(blockIdx.x: int32, (nullptr), "ThreadIndex", "blockIdx.x")] "thread_extent" = 16;
attr [compute_3: Pointer(float32)] "storage_scope" = "local";
allocate(compute_3, float32, [14]);
attr [pad_temp.shared: Pointer(float32)] "storage_scope" = "shared";
allocate(pad_temp.shared, float32, [1296]);
attr [kernel.shared: Pointer(float32)] "storage_scope" = "shared";
allocate(kernel.shared, float32, [4608]);
attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
compute_3[0] = 0f32
compute_3[7] = 0f32
compute_3[1] = 0f32
compute_3[8] = 0f32
compute_3[2] = 0f32
compute_3[9] = 0f32
compute_3[3] = 0f32
compute_3[10] = 0f32
compute_3[4] = 0f32
compute_3[11] = 0f32
compute_3[5] = 0f32
compute_3[12] = 0f32
compute_3[6] = 0f32
compute_3[13] = 0f32
for (rc.outer.outer: int32, 0, 32) {
attr [IterVar(threadIdx.x_1: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[threadIdx.x_1] = @tir.if_then_else(((((9 <= floormod(threadIdx.x_1, 81)) && (floormod(threadIdx.x_1, 81) < 72)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv(threadIdx.x_1, 81)*49)) + (floordiv(floormod(threadIdx.x_1, 81), 9)*7)) + floormod(threadIdx.x_1, 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 112)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 31), 81)) && (floormod((threadIdx.x_1 + 31), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 112), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 31), 81), 9)*7)) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 224)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 62), 81)) && (floormod((threadIdx.x_1 + 62), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 224), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 62), 81), 9)*7)) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 336)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 12), 81)) && (floormod((threadIdx.x_1 + 12), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 3), 9))) && (floormod((threadIdx.x_1 + 3), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 336), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 12), 81), 9)*7)) + floormod((threadIdx.x_1 + 3), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 448)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 43), 81)) && (floormod((threadIdx.x_1 + 43), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 448), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 43), 81), 9)*7)) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 560)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 74), 81)) && (floormod((threadIdx.x_1 + 74), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 2), 9))) && (floormod((threadIdx.x_1 + 2), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 560), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 74), 81), 9)*7)) + floormod((threadIdx.x_1 + 2), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 672)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 24), 81)) && (floormod((threadIdx.x_1 + 24), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 6), 9))) && (floormod((threadIdx.x_1 + 6), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 672), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 24), 81), 9)*7)) + floormod((threadIdx.x_1 + 6), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 784)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 55), 81)) && (floormod((threadIdx.x_1 + 55), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 1), 9))) && (floormod((threadIdx.x_1 + 1), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 784), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 55), 81), 9)*7)) + floormod((threadIdx.x_1 + 1), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 896)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 5), 81)) && (floormod((threadIdx.x_1 + 5), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 5), 9))) && (floormod((threadIdx.x_1 + 5), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 896), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 5), 81), 9)*7)) + floormod((threadIdx.x_1 + 5), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 1008)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 36), 81)) && (floormod((threadIdx.x_1 + 36), 81) < 72)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 1008), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 36), 81), 9)*7)) + floormod(threadIdx.x_1, 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
pad_temp.shared[(threadIdx.x_1 + 1120)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 67), 81)) && (floormod((threadIdx.x_1 + 67), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 1120), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 67), 81), 9)*7)) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
if @tir.likely((threadIdx.x_1 < 64), dtype=bool) {
pad_temp.shared[(threadIdx.x_1 + 1232)] = @tir.if_then_else((((floormod((threadIdx.x_1 + 17), 81) < 72) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), (float32*)data_2[(((((rc.outer.outer*784) + (floordiv((threadIdx.x_1 + 1232), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 17), 81), 9)*7)) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
}
attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[(threadIdx.x_2*4)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 36)*4608)) + (rc.outer.outer*144)) + (floormod(threadIdx.x_2, 36)*4))]
kernel.shared[((threadIdx.x_2*4) + 1)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 1), 144))]
kernel.shared[((threadIdx.x_2*4) + 2)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 2), 144))]
kernel.shared[((threadIdx.x_2*4) + 3)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 3), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 448)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 448), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 16), 144))]
kernel.shared[((threadIdx.x_2*4) + 449)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 449), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 17), 144))]
kernel.shared[((threadIdx.x_2*4) + 450)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 450), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 18), 144))]
kernel.shared[((threadIdx.x_2*4) + 451)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 451), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 19), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 896)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 896), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 32), 144))]
kernel.shared[((threadIdx.x_2*4) + 897)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 897), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 33), 144))]
kernel.shared[((threadIdx.x_2*4) + 898)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 898), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 34), 144))]
kernel.shared[((threadIdx.x_2*4) + 899)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 899), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 35), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 1344)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1344), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 48), 144))]
kernel.shared[((threadIdx.x_2*4) + 1345)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1345), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 49), 144))]
kernel.shared[((threadIdx.x_2*4) + 1346)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1346), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 50), 144))]
kernel.shared[((threadIdx.x_2*4) + 1347)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1347), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 51), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 1792)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1792), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 64), 144))]
kernel.shared[((threadIdx.x_2*4) + 1793)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1793), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 65), 144))]
kernel.shared[((threadIdx.x_2*4) + 1794)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1794), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 66), 144))]
kernel.shared[((threadIdx.x_2*4) + 1795)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 1795), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 67), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 2240)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2240), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 80), 144))]
kernel.shared[((threadIdx.x_2*4) + 2241)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2241), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 81), 144))]
kernel.shared[((threadIdx.x_2*4) + 2242)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2242), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 82), 144))]
kernel.shared[((threadIdx.x_2*4) + 2243)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2243), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 83), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 2688)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2688), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 96), 144))]
kernel.shared[((threadIdx.x_2*4) + 2689)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2689), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 97), 144))]
kernel.shared[((threadIdx.x_2*4) + 2690)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2690), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 98), 144))]
kernel.shared[((threadIdx.x_2*4) + 2691)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 2691), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 99), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 3136)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3136), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 112), 144))]
kernel.shared[((threadIdx.x_2*4) + 3137)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3137), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 113), 144))]
kernel.shared[((threadIdx.x_2*4) + 3138)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3138), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 114), 144))]
kernel.shared[((threadIdx.x_2*4) + 3139)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3139), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 115), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 3584)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3584), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 128), 144))]
kernel.shared[((threadIdx.x_2*4) + 3585)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3585), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 129), 144))]
kernel.shared[((threadIdx.x_2*4) + 3586)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3586), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 130), 144))]
kernel.shared[((threadIdx.x_2*4) + 3587)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 3587), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 131), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
kernel.shared[((threadIdx.x_2*4) + 4032)] = (float32*)kernel_2[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2*4), 144)*4608)) + (rc.outer.outer*144)) + (floormod(threadIdx.x_2, 36)*4)) + 129024)]
kernel.shared[((threadIdx.x_2*4) + 4033)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4033), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 1), 144))]
kernel.shared[((threadIdx.x_2*4) + 4034)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4034), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 2), 144))]
kernel.shared[((threadIdx.x_2*4) + 4035)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4035), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 3), 144))]
}
attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
if @tir.likely((threadIdx.x_2 < 32), dtype=bool) {
kernel.shared[((threadIdx.x_2*4) + 4480)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4480), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 16), 144))]
}
if @tir.likely(((threadIdx.x_2*4) < 127), dtype=bool) {
if @tir.likely((threadIdx.x_2 < 32), dtype=bool) {
kernel.shared[((threadIdx.x_2*4) + 4481)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4481), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 17), 144))]
}
}
if @tir.likely(((threadIdx.x_2*4) < 126), dtype=bool) {
if @tir.likely((threadIdx.x_2 < 32), dtype=bool) {
kernel.shared[((threadIdx.x_2*4) + 4482)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4482), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 18), 144))]
}
}
if @tir.likely(((threadIdx.x_2*4) < 125), dtype=bool) {
if @tir.likely((threadIdx.x_2 < 32), dtype=bool) {
kernel.shared[((threadIdx.x_2*4) + 4483)] = (float32*)kernel_2[((((blockIdx.x*147456) + (floordiv(((threadIdx.x_2*4) + 4483), 144)*4608)) + (rc.outer.outer*144)) + floormod(((threadIdx.x_2*4) + 19), 144))]
}
}
}
for (rc.outer.inner: int32, 0, 4) {
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9))]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9))]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 1)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 1)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36))]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2304)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 1)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 1)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 7)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 1)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 7)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2305)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 2)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 3)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 4)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 5)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 6)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 7)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 7)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 8)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 8)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2306)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 81)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 81)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 82)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 82)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 9)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2313)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 82)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 82)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 88)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 10)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 88)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2314)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 83)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 84)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 85)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 86)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 87)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 88)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 88)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 89)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 11)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 89)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2315)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 162)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 162)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 163)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 163)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 18)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2322)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 163)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 163)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 169)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 19)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 169)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2323)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 164)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 165)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 166)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 167)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 168)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 169)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 169)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 170)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 20)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 170)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2324)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 243)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 243)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 244)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 244)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 27)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2331)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 244)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 244)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 250)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 28)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 250)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2332)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 245)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 246)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 247)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 248)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 249)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 250)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 250)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 251)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 29)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 251)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2333)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 9)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 9)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 10)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 10)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 3)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2307)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 10)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 10)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 16)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 4)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 16)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2308)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 11)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 12)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 13)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 14)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 15)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 16)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 16)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 17)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 5)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 17)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2309)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 90)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 90)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 91)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 91)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 12)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2316)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 91)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 91)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 97)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 13)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 97)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2317)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 92)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 93)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 94)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 95)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 96)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 97)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 97)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 98)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 14)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 98)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2318)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 171)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 171)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 172)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 172)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 21)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2325)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 172)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 172)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 178)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 22)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 178)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2326)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 173)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 174)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 175)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 176)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 177)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 178)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 178)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 179)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 23)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 179)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2327)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 252)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 252)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 253)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 253)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 30)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2334)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 253)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 253)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 259)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 31)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 259)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2335)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 254)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 255)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 256)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 257)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 258)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 259)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 259)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 260)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 32)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 260)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2336)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 18)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 18)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 19)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 19)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 6)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2310)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 19)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 19)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 25)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 7)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 25)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2311)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 20)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 21)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 22)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 23)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 24)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 25)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 25)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 26)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 8)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 26)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2312)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 99)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 99)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 100)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 100)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 15)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2319)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 100)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 100)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 106)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 16)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 106)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2320)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 101)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 102)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 103)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 104)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 105)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 106)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 106)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 107)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 17)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 107)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2321)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 180)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 180)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 181)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 181)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 24)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2328)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 181)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 181)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 187)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 25)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 187)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2329)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 182)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 183)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 184)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 185)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 186)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 187)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 187)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 188)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 26)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 188)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2330)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 261)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 261)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 262)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 262)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 33)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2337)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 262)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 262)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 268)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 34)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 268)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2338)]))
compute_3[0] = ((float32*)compute_3[0] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[7] = ((float32*)compute_3[7] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 263)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[1] = ((float32*)compute_3[1] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[8] = ((float32*)compute_3[8] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 264)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[2] = ((float32*)compute_3[2] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[9] = ((float32*)compute_3[9] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 265)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[3] = ((float32*)compute_3[3] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[10] = ((float32*)compute_3[10] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 266)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[4] = ((float32*)compute_3[4] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[11] = ((float32*)compute_3[11] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 267)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[5] = ((float32*)compute_3[5] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 268)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[12] = ((float32*)compute_3[12] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 268)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
compute_3[6] = ((float32*)compute_3[6] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 269)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 35)]))
compute_3[13] = ((float32*)compute_3[13] + ((float32*)pad_temp.shared[(((rc.outer.inner*324) + (floormod(threadIdx.x, 7)*9)) + 269)]*(float32*)kernel.shared[(((floordiv(threadIdx.x, 7)*144) + (rc.outer.inner*36)) + 2339)]))
}
}
for (i3.inner: int32, 0, 7) {
compute_2[(((blockIdx.x*1568) + (threadIdx.x*7)) + i3.inner)] = max(((float32*)compute_3[i3.inner] + (float32*)bias_2[((blockIdx.x*32) + floordiv(threadIdx.x, 7))]), 0f32)
compute_2[((((blockIdx.x*1568) + (threadIdx.x*7)) + i3.inner) + 784)] = max(((float32*)compute_3[(i3.inner + 7)] + (float32*)bias_2[(((blockIdx.x*32) + floordiv(threadIdx.x, 7)) + 16)]), 0f32)
}
}
}
检查正确性并评估性能
构建二进制文件并检查其正确性和性能。
func = tvm.build(sch, args, target)
# Check correctness
data_np = np.random.uniform(size=(N, CI, H, W)).astype(np.float32)
weight_np = np.random.uniform(size=(CO, CI, KH, KW)).astype(np.float32)
bias_np = np.random.uniform(size=(1, CO, 1, 1)).astype(np.float32)
conv_np = conv2d_nchw_python(data_np, weight_np, strides, padding)
out_np = np.maximum(conv_np + bias_np, 0.0)
ctx = tvm.gpu()
data_tvm = tvm.nd.array(data_np, ctx=ctx)
weight_tvm = tvm.nd.array(weight_np, ctx=ctx)
bias_tvm = tvm.nd.array(bias_np, ctx=ctx)
out_tvm = tvm.nd.empty(out_np.shape, ctx=ctx)
func(data_tvm, weight_tvm, bias_tvm, out_tvm)
# Check results
np.testing.assert_allclose(out_np, out_tvm.asnumpy(), rtol=1e-3)
# Evaluate execution time
evaluator = func.time_evaluator(func.entry_name, ctx, min_repeat_ms=500)
print(
"Execution time of this operator: %.3f ms"
% (np.median(evaluator(data_tvm, weight_tvm, bias_tvm, out_tvm).results) * 1000)
)
输出:
Execution time of this operator: 0.184 ms
使用记录文件
搜索期间,所有测量记录都将转储到记录文件“ conv2d.json”中。测量记录可用于重新应用搜索结果,继续搜索以及执行其它分析。
这是一个示例,其中从文件加载最佳调度,打印等效的python调度API和CUDA源代码。它们可用于调试和学习自动调度程序的行为。
print("Equivalent python schedule:")
print(task.print_best(log_file, print_mode="schedule"))
print("CUDA source code:")
print(task.print_best(log_file, print_mode="cuda"))
输出:
Equivalent python schedule:
pad_temp_i0, pad_temp_i1, pad_temp_i2, pad_temp_i3 = tuple(pad_temp.op.axis) + tuple(pad_temp.op.reduce_axis)
compute_nn, compute_ff, compute_yy, compute_xx, compute_rc, compute_ry, compute_rx = tuple(compute.op.axis) + tuple(compute.op.reduce_axis)
T_add_ax0, T_add_ax1, T_add_ax2, T_add_ax3 = tuple(T_add.op.axis) + tuple(T_add.op.reduce_axis)
compute_i0, compute_i1, compute_i2, compute_i3 = tuple(compute.op.axis) + tuple(compute.op.reduce_axis)
s[T_add].compute_inline()
compute_nn_o_i, compute_nn_i = s[compute].split(compute_nn, factor=1)
compute_nn_o_o_i, compute_nn_o_i = s[compute].split(compute_nn_o_i, factor=1)
compute_nn_o_o_o_i, compute_nn_o_o_i = s[compute].split(compute_nn_o_o_i, factor=1)
compute_nn_o_o_o_o, compute_nn_o_o_o_i = s[compute].split(compute_nn_o_o_o_i, factor=1)
compute_ff_o_i, compute_ff_i = s[compute].split(compute_ff, factor=1)
compute_ff_o_o_i, compute_ff_o_i = s[compute].split(compute_ff_o_i, factor=1)
compute_ff_o_o_o_i, compute_ff_o_o_i = s[compute].split(compute_ff_o_o_i, factor=16)
compute_ff_o_o_o_o, compute_ff_o_o_o_i = s[compute].split(compute_ff_o_o_o_i, factor=2)
compute_yy_o_i, compute_yy_i = s[compute].split(compute_yy, factor=1)
compute_yy_o_o_i, compute_yy_o_i = s[compute].split(compute_yy_o_i, factor=1)
compute_yy_o_o_o_i, compute_yy_o_o_i = s[compute].split(compute_yy_o_o_i, factor=7)
compute_yy_o_o_o_o, compute_yy_o_o_o_i = s[compute].split(compute_yy_o_o_o_i, factor=1)
compute_xx_o_i, compute_xx_i = s[compute].split(compute_xx, factor=7)
compute_xx_o_o_i, compute_xx_o_i = s[compute].split(compute_xx_o_i, factor=1)
compute_xx_o_o_o_i, compute_xx_o_o_i = s[compute].split(compute_xx_o_o_i, factor=1)
compute_xx_o_o_o_o, compute_xx_o_o_o_i = s[compute].split(compute_xx_o_o_o_i, factor=1)
compute_rc_o_i, compute_rc_i = s[compute].split(compute_rc, factor=4)
compute_rc_o_o, compute_rc_o_i = s[compute].split(compute_rc_o_i, factor=4)
compute_ry_o_i, compute_ry_i = s[compute].split(compute_ry, factor=1)
compute_ry_o_o, compute_ry_o_i = s[compute].split(compute_ry_o_i, factor=3)
compute_rx_o_i, compute_rx_i = s[compute].split(compute_rx, factor=3)
compute_rx_o_o, compute_rx_o_i = s[compute].split(compute_rx_o_i, factor=1)
s[compute].reorder(compute_nn_o_o_o_o, compute_ff_o_o_o_o, compute_yy_o_o_o_o, compute_xx_o_o_o_o, compute_nn_o_o_o_i, compute_ff_o_o_o_i, compute_yy_o_o_o_i, compute_xx_o_o_o_i, compute_nn_o_o_i, compute_ff_o_o_i, compute_yy_o_o_i, compute_xx_o_o_i, compute_rc_o_o, compute_ry_o_o, compute_rx_o_o, compute_rc_o_i, compute_ry_o_i, compute_rx_o_i, compute_nn_o_i, compute_ff_o_i, compute_yy_o_i, compute_xx_o_i, compute_rc_i, compute_ry_i, compute_rx_i, compute_nn_i, compute_ff_i, compute_yy_i, compute_xx_i)
compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=1)
compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=16)
compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
compute_i2_o_o_o, compute_i2_o_o_i = s[compute].split(compute_i2_o_o_i, factor=1)
compute_i3_o_i, compute_i3_i = s[compute].split(compute_i3, factor=7)
compute_i3_o_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=1)
compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=1)
s[compute].reorder(compute_i0_o_o_o, compute_i1_o_o_o, compute_i2_o_o_o, compute_i3_o_o_o, compute_i0_o_o_i, compute_i1_o_o_i, compute_i2_o_o_i, compute_i3_o_o_i, compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i, compute_i0_i, compute_i1_i, compute_i2_i, compute_i3_i)
s[compute].compute_at(s[compute], compute_i3_o_i)
kernel_shared = s.cache_read(kernel, "shared", [compute])
kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3 = tuple(kernel_shared.op.axis)
s[kernel_shared].compute_at(s[compute], compute_rx_o_o)
pad_temp_shared = s.cache_read(pad_temp, "shared", [compute])
pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3 = tuple(pad_temp_shared.op.axis)
s[pad_temp_shared].compute_at(s[compute], compute_rx_o_o)
s[pad_temp].compute_inline()
compute_i0_o_o_o_i1_o_o_o_fused_i2_o_o_o_fused_i3_o_o_o_fused = s[compute].fuse(compute_i0_o_o_o, compute_i1_o_o_o, compute_i2_o_o_o, compute_i3_o_o_o)
s[compute].bind(compute_i0_o_o_o_i1_o_o_o_fused_i2_o_o_o_fused_i3_o_o_o_fused, te.thread_axis("blockIdx.x"))
compute_i0_o_o_i_i1_o_o_i_fused_i2_o_o_i_fused_i3_o_o_i_fused = s[compute].fuse(compute_i0_o_o_i, compute_i1_o_o_i, compute_i2_o_o_i, compute_i3_o_o_i)
s[compute].bind(compute_i0_o_o_i_i1_o_o_i_fused_i2_o_o_i_fused_i3_o_o_i_fused, te.thread_axis("vthread"))
compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused = s[compute].fuse(compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i)
s[compute].bind(compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused, te.thread_axis("threadIdx.x"))
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[kernel_shared].fuse(kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3)
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=4)
s[kernel_shared].vectorize(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=112)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[pad_temp_shared].fuse(pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=112)
s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
s[compute].pragma(compute_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[compute].pragma(compute_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
#ifdef _WIN32
using uint = unsigned int;
using uchar = unsigned char;
using ushort = unsigned short;
using int64_t = long long;
using uint64_t = unsigned long long;
#else
#define uint unsigned int
#define uchar unsigned char
#define ushort unsigned short
#define int64_t long
#define uint64_t ulong
#endif
extern "C" __global__ void default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float compute1[14];
__shared__ float pad_temp_shared[1296];
__shared__ float kernel_shared[4608];
compute1[(0)] = 0.000000e+00f;
compute1[(7)] = 0.000000e+00f;
compute1[(1)] = 0.000000e+00f;
compute1[(8)] = 0.000000e+00f;
compute1[(2)] = 0.000000e+00f;
compute1[(9)] = 0.000000e+00f;
compute1[(3)] = 0.000000e+00f;
compute1[(10)] = 0.000000e+00f;
compute1[(4)] = 0.000000e+00f;
compute1[(11)] = 0.000000e+00f;
compute1[(5)] = 0.000000e+00f;
compute1[(12)] = 0.000000e+00f;
compute1[(6)] = 0.000000e+00f;
compute1[(13)] = 0.000000e+00f;
for (int rc_outer_outer = 0; rc_outer_outer < 32; ++rc_outer_outer) {
__syncthreads();
pad_temp_shared[(((int)threadIdx.x))] = (((((9 <= (((int)threadIdx.x) % 81)) && ((((int)threadIdx.x) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + ((((int)threadIdx.x) / 81) * 49)) + (((((int)threadIdx.x) % 81) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 112))] = (((((9 <= ((((int)threadIdx.x) + 31) % 81)) && (((((int)threadIdx.x) + 31) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 112) / 81) * 49)) + ((((((int)threadIdx.x) + 31) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 224))] = (((((9 <= ((((int)threadIdx.x) + 62) % 81)) && (((((int)threadIdx.x) + 62) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 224) / 81) * 49)) + ((((((int)threadIdx.x) + 62) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 336))] = (((((9 <= ((((int)threadIdx.x) + 12) % 81)) && (((((int)threadIdx.x) + 12) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 336) / 81) * 49)) + ((((((int)threadIdx.x) + 12) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 448))] = (((((9 <= ((((int)threadIdx.x) + 43) % 81)) && (((((int)threadIdx.x) + 43) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 448) / 81) * 49)) + ((((((int)threadIdx.x) + 43) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 560))] = (((((9 <= ((((int)threadIdx.x) + 74) % 81)) && (((((int)threadIdx.x) + 74) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 560) / 81) * 49)) + ((((((int)threadIdx.x) + 74) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 672))] = (((((9 <= ((((int)threadIdx.x) + 24) % 81)) && (((((int)threadIdx.x) + 24) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 672) / 81) * 49)) + ((((((int)threadIdx.x) + 24) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 784))] = (((((9 <= ((((int)threadIdx.x) + 55) % 81)) && (((((int)threadIdx.x) + 55) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 784) / 81) * 49)) + ((((((int)threadIdx.x) + 55) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 896))] = (((((9 <= ((((int)threadIdx.x) + 5) % 81)) && (((((int)threadIdx.x) + 5) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 896) / 81) * 49)) + ((((((int)threadIdx.x) + 5) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 1008))] = (((((9 <= ((((int)threadIdx.x) + 36) % 81)) && (((((int)threadIdx.x) + 36) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 1008) / 81) * 49)) + ((((((int)threadIdx.x) + 36) % 81) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8))] : 0.000000e+00f);
pad_temp_shared[((((int)threadIdx.x) + 1120))] = (((((9 <= ((((int)threadIdx.x) + 67) % 81)) && (((((int)threadIdx.x) + 67) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 1120) / 81) * 49)) + ((((((int)threadIdx.x) + 67) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8))] : 0.000000e+00f);
if (((int)threadIdx.x) < 64) {
pad_temp_shared[((((int)threadIdx.x) + 1232))] = ((((((int)threadIdx.x) < 55) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 1232) / 81) * 49)) + (((((int)threadIdx.x) + 17) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8))] : 0.000000e+00f);
}
kernel_shared[((((int)threadIdx.x) * 4))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 36) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) % 36) * 4)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 1) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 2) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 3) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 448))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 448) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 16) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 449))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 449) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 17) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 450))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 450) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 18) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 451))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 451) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 19) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 896))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 896) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 32) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 897))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 897) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 33) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 898))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 898) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 34) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 899))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 899) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 35) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1344))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1344) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 48) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1345))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1345) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 49) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1346))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1346) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 50) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1347))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1347) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 51) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1792))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1792) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 64) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1793))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1793) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 65) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1794))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1794) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 66) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 1795))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 1795) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 67) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2240))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2240) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 80) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2241))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2241) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 81) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2242))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2242) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 82) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2243))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2243) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 83) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2688))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2688) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 96) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2689))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2689) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 97) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2690))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2690) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 98) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 2691))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 2691) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 99) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3136))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3136) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 112) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3137))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3137) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 113) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3138))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3138) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 114) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3139))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3139) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 115) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3584))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3584) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 128) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3585))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3585) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 129) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3586))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3586) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 130) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 3587))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 3587) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 131) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 4032))] = kernel[((((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 36) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) % 36) * 4)) + 129024))];
kernel_shared[(((((int)threadIdx.x) * 4) + 4033))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4033) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 1) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 4034))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4034) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 2) % 144)))];
kernel_shared[(((((int)threadIdx.x) * 4) + 4035))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4035) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) * 4) + 3) % 144)))];
if (((int)threadIdx.x) < 32) {
kernel_shared[(((((int)threadIdx.x) * 4) + 4480))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4480) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) * 4) + 16)))];
}
if (((int)threadIdx.x) < 32) {
kernel_shared[(((((int)threadIdx.x) * 4) + 4481))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4481) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) * 4) + 17)))];
}
if (((int)threadIdx.x) < 32) {
kernel_shared[(((((int)threadIdx.x) * 4) + 4482))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4482) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) * 4) + 18)))];
}
if (((int)threadIdx.x) < 32) {
kernel_shared[(((((int)threadIdx.x) * 4) + 4483))] = kernel[(((((((int)blockIdx.x) * 147456) + ((((((int)threadIdx.x) * 4) + 4483) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((int)threadIdx.x) * 4) + 19)))];
}
__syncthreads();
for (int rc_outer_inner = 0; rc_outer_inner < 4; ++rc_outer_inner) {
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[(((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[(((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 1))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 1))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2304))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 1))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 1))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 1))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2305))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 2))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 3))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 4))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 5))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 6))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 8))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 8))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2306))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 81))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 81))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 82))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 82))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 9))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2313))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 82))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 82))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 88))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 10))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 88))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2314))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 83))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 84))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 85))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 86))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 87))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 88))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 88))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 89))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 11))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 89))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2315))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 162))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 162))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 163))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 163))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 18))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2322))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 163))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 163))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 169))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 19))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 169))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2323))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 164))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 165))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 166))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 167))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 168))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 169))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 169))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 170))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 20))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 170))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2324))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 243))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 243))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 244))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 244))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 27))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2331))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 244))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 244))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 250))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 28))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 250))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2332))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 245))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 246))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 247))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 248))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 249))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 250))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 250))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 251))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 29))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 251))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2333))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 9))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 9))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 10))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 10))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 3))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2307))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 10))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 10))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 16))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 4))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 16))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2308))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 11))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 12))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 13))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 14))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 15))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 16))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 16))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 17))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 5))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 17))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2309))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 90))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 90))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 91))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 91))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 12))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2316))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 91))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 91))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 97))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 13))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 97))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2317))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 92))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 93))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 94))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 95))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 96))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 97))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 97))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 98))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 14))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 98))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2318))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 171))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 171))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 172))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 172))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 21))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2325))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 172))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 172))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 178))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 22))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 178))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2326))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 173))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 174))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 175))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 176))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 177))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 178))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 178))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 179))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 23))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 179))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2327))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 252))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 252))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 253))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 253))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 30))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2334))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 253))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 253))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 259))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 31))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 259))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2335))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 254))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 255))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 256))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 257))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 258))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 259))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 259))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 260))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 32))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 260))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2336))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 18))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 18))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 19))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 19))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 6))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2310))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 19))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 19))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 25))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 7))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 25))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2311))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 20))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 21))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 22))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 23))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 24))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 25))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 25))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 26))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 8))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 26))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2312))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 99))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 99))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 100))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 100))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 15))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2319))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 100))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 100))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 106))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 16))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 106))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2320))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 101))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 102))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 103))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 104))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 105))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 106))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 106))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 107))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 17))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 107))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2321))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 180))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 180))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 181))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 181))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 24))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2328))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 181))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 181))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 187))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 25))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 187))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2329))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 182))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 183))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 184))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 185))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 186))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 187))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 187))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 188))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 26))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 188))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2330))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 261))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 261))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 262))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 262))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 33))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2337))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 262))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 262))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 268))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 34))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 268))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2338))]));
compute1[(0)] = (compute1[(0)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(7)] = (compute1[(7)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 263))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(1)] = (compute1[(1)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(8)] = (compute1[(8)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 264))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(2)] = (compute1[(2)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(9)] = (compute1[(9)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 265))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(3)] = (compute1[(3)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(10)] = (compute1[(10)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 266))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(4)] = (compute1[(4)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(11)] = (compute1[(11)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 267))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(5)] = (compute1[(5)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 268))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(12)] = (compute1[(12)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 268))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
compute1[(6)] = (compute1[(6)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 269))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 35))]));
compute1[(13)] = (compute1[(13)] + (pad_temp_shared[((((rc_outer_inner * 324) + ((((int)threadIdx.x) % 7) * 9)) + 269))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 144) + (rc_outer_inner * 36)) + 2339))]));
}
}
for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
compute[((((((int)blockIdx.x) * 1568) + (((int)threadIdx.x) * 7)) + i3_inner))] = max((compute1[(i3_inner)] + bias[(((((int)blockIdx.x) * 32) + (((int)threadIdx.x) / 7)))]), 0.000000e+00f);
compute[(((((((int)blockIdx.x) * 1568) + (((int)threadIdx.x) * 7)) + i3_inner) + 784))] = max((compute1[((i3_inner + 7))] + bias[((((((int)blockIdx.x) * 32) + (((int)threadIdx.x) / 7)) + 16))]), 0.000000e+00f);
}
}
一个更复杂的示例是继续搜索。在这种情况下,需要自己创建搜索策略和成本模型,使用日志文件恢复搜索策略和成本模型的状态。在下面的示例中,恢复状态并进行5次以上的试用。
def resume_search(task, log_file):
print("Resume search:")
cost_model = auto_scheduler.XGBModel()
cost_model.update_from_file(log_file)
search_policy = auto_scheduler.SketchPolicy(
task, cost_model, init_search_callbacks=[auto_scheduler.PreloadMeasuredStates(log_file)]
)
measure_ctx = auto_scheduler.LocalRPCMeasureContext(min_repeat_ms=300)
tune_option = auto_scheduler.TuningOptions(
num_measure_trials=5,
runner=measure_ctx.runner,
measure_callbacks=[auto_scheduler.RecordToFile(log_file)],
)
task.tune(tune_option, search_policy=search_policy)
# Kill the measurement process
del measure_ctx
resume_search(task, log_file)
输出:
Resume search:
Get devices for measurement successfully!
脚本的总运行时间:(1分钟37.345秒)
GPU自动调度卷积层的更多相关文章
- NVIDIA GPU自动调度神经网络
NVIDIA GPU自动调度神经网络 对特定设备和工作负载进行自动调整对于获得最佳性能至关重要.这是有关如何使用自动调度器为NVIDIA GPU调整整个神经网络. 为了自动调整神经网络,将网络划分为小 ...
- 自动调度GPU的卷积层
自动调度GPU的卷积层 这是有关如何对GPU使用自动调度程序的文档. 与依靠手动模板定义搜索空间的基于模板的autotvm不同,自动调度程序不需要任何模板.用户只需要编写计算声明,而无需任何调度命令或 ...
- NVIDIA GPU的神经网络自动调度
NVIDIA GPU的神经网络自动调度 针对特定设备和工作负载的自动调整对于获得最佳性能至关重要.这是一个关于如何使用自动调度器为NVIDIA GPU调整整个神经网络的资料. 为了自动调整一个神经网络 ...
- CPU的自动调度矩阵乘法
CPU的自动调度矩阵乘法 这是一个有关如何对CPU使用自动调度程序的文档. 与依靠手动模板定义搜索空间的基于模板的autotvm不同,自动调度程序不需要任何模板.用户只需要编写计算声明,而无需任何调度 ...
- TVM自动调度器
TVM自动调度器 随着模型大小,算子多样性和硬件异构性的不断增长,优化深度神经网络的执行速度非常困难.从计算的角度来看,深度神经网络只是张量计算的一层又一层.这些张量计算(例如matmul和conv2 ...
- tensorflow 1.0 学习:卷积层
在tf1.0中,对卷积层重新进行了封装,比原来版本的卷积层有了很大的简化. 一.旧版本(1.0以下)的卷积函数:tf.nn.conv2d conv2d( input, filter, strides, ...
- Keras深度神经网络算法模型构建【输入层、卷积层、池化层】
一.输入层 1.用途 构建深度神经网络输入层,确定输入数据的类型和样式. 2.应用代码 input_data = Input(name='the_input', shape=(1600, 200, 1 ...
- ARM CPU自动调度神经网络
ARM CPU自动调度神经网络 对特定设备和工作负载进行自动调度,对于获得最佳性能至关重要.通过RPC使用自动调度器为ARM CPU调度整个神经网络. 为了自动调度神经网络,将网络划分为小的子图,进行 ...
- 为x86 CPU自动调度神经网络
为x86 CPU自动调度神经网络 对特定设备和工作负载进行自动调试对于获得最佳性能至关重要.这是有关如何使用自动调度器为x86 CPU调试整个神经网络的文档. 为了自动调试神经网络,将网络划分为小的子 ...
随机推荐
- Git使用指导
Git使用指导 目录结构: 一.版本控制 二.Git的历史 三.Git环境配置 四.GIt项目构建 五.Git基本理论 六.GIt文件操作 七.使用码云/GitHub 八.IDEA中继承Git 九.G ...
- Mysql 8.0安装
1. 下载安装包至/usr/local目录下 下载地址:https://cdn.mysql.com/Downloads/MySQL-8.0/mysql-8.0.16-el7-x86_64.tar.gz ...
- hdu1686 最大匹配次数 KMP
题意: 给你两个串,问你串a在串b中出现了多少次. 思路: 直接匹配,KMP时匹配到匹配串的最后一个的时候不用跳出,直接匹配就行了,最后一个'/0'不会和目标串匹配,所以经过ne ...
- WAF、流控设备、堡垒机
目录 WAF 流控设备 堡垒机 WAF WAF(Web Application Firewall):web防火墙,WAF是通过执行一系列针对HTTP/HTTPS的安全策略来专门对web应用提供保护的一 ...
- (邹博ML)数学分析与概率论
机器学习入门 深度学习和机器学习? 深度学习在某种意义上可以认为是机器学习的一个分支,只是这个分支非常全面且重要,以至于可以单独作为一门学科来进行研究. 回忆知识 求解S. 对数函数的上升速度 我们使 ...
- PHP基础-PHP中预定义的超全局数组
预定义数组: 自动全局变量---超全局数组 1. 包含了来自WEB服务器,客户端,运行环境和用户输入的数据 2. 这些数组比较特别 3. 全局范围内自动生效,都可以直接使用这些数组 4. 用户不能自定 ...
- 注解配置定时器Scheduling
注解配置定时器配置 package com.demo; import org.springframework.context.annotation.Configuration; import org. ...
- 【Tavern】1.tavern测试框架介绍
目录 1. Tavern 和 Postman 2. 安装Python 3. 安装Tavern 4. 第一个测试用例 1. Tavern 和 Postman Tavern是一个用Python实现的接口自 ...
- (10)MySQL进阶篇SQL优化(InnoDB锁-间隙锁)
1.概述 当我们用范围条件而不是相等条件检索数据,并请求共享或排他锁时,InnoDB会给符合条件的已有数据记录的索引项加锁:对于键值在条件范围内但并不存在的记录,叫做"间隙(GAP)&quo ...
- Js的变量、作用域与内存
变量.作用域与内存 1 .原始值与引用值 Undefined.Null.Boolean.Number. String和Symbol.保存原始值的变量是按值(by value)访问的 引用值是保存在内存 ...