leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

CUTLASS 2.7 (#318)

Browse Source 
CUTLASS 2.7

Mainloop fusion for GEMM: summation over A or B
Strided DGRAD (optimized iterators)
Half-precision GELU_taylor activation functions
Use these when accumulation and epilogue compute types are all cutlass::half_t
Tuning and bug fixes to fused GEMM + GEMM example
Support for smaller than 128b aligned Convolutions: see examples
Caching of results to accelerate Convolution unit tests
Can be enabled or disabled by running cmake .. -DCUTLASS_TEST_ENABLE_CACHED_RESULTS=OFF
Corrections and bug fixes reported by the CUTLASS community
Thank you for filing these issues!

authored-by: Haicheng Wu haichengw@nvidia.com, Manish Gupta manigupta@nvidia.com, Dustyn Blasig dblasig@nvidia.com, Andrew Kerr akerr@nvidia.com
This commit is contained in:
Manish Gupta
2021-09-20 11:02:22 -07:00
committed by GitHub
parent 9ac255863f
commit 2e07c4cc2f
62 changed files with 5611 additions and 186 deletions
Download Patch File Download Diff File Expand all files Collapse all files

60
test/unit/conv/device/conv2d_problems.h
View File

@@ -182,7 +182,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{1, 7, 8, minimum_channel_size}, // input size (NHWC)
{8, 3, 3, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -190,7 +190,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{1, 7, 9, minimum_channel_size}, // input size (NHWC)
{8, 4, 4, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -198,7 +198,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{2, 7, 9, minimum_channel_size}, // input size (NHWC)
{8, 5, 5, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -206,7 +206,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{3, 7, 9, minimum_channel_size}, // input size (NHWC)
{8, 6, 5, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -214,7 +214,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{3, 7, 9, minimum_channel_size}, // input size (NHWC)
{8, 6, 6, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -222,7 +222,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{3, 7, 9, minimum_channel_size}, // input size (NHWC)
{8, 7, 7, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -234,7 +234,7 @@ struct TestbedConv2dProblemSizes {
// C < CTA::K and non-multiples of CTA::K. Typical CTA::K = {32, 64}
////////////////////////////////////////////////////////////////////////////////////////////
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 11, 11, minimum_channel_size}, // input size (NHWC)
{1, 11, 7, minimum_channel_size}, // input size (NHWC)
{8, 1, 1, minimum_channel_size}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -242,7 +242,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 11, 11, minimum_channel_size}, // input size (NHWC)
{1, 11, 7, minimum_channel_size}, // input size (NHWC)
{8, 3, 3, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -250,7 +250,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 13, 13, minimum_channel_size}, // input size (NHWC)
{1, 13, 11, minimum_channel_size}, // input size (NHWC)
{8, 1, 1, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -258,40 +258,40 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, minimum_channel_size}, // input size (NHWC)
{8, 2, 2, minimum_channel_size}, // filter size (KRSC)
{1, 17, 19, minimum_channel_size}, // input size (NHWC)
{16, 2, 2, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 5, 5, minimum_channel_size}, // input size (NHWC)
{8, 3, 3, minimum_channel_size}, // filter size (KRSC)
{1, 23, 5, minimum_channel_size}, // input size (NHWC)
{16, 3, 3, minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, 8}, // input size (NHWC)
{8, 3, 3, 8}, // filter size (KRSC)
{1, 13, 17, 8}, // input size (NHWC)
{24, 3, 3, 8}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 8, 8, 8}, // input size (NHWC)
{8, 3, 3, 8}, // filter size (KRSC)
{1, 23, 21, 8}, // input size (NHWC)
{24, 3, 3, 8}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{3, 3}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 16, 16, 8}, // input size (NHWC)
{8, 3, 3, 8}, // filter size (KRSC)
{1, 20, 24, 8}, // input size (NHWC)
{40, 3, 3, 8}, // filter size (KRSC)
{3, 3, 3, 3}, // padding (pad_h, _, pad_w, _)
{3, 3}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
@@ -352,7 +352,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 16, 16, 64 + minimum_channel_size}, // input size (NHWC)
{1, 16, 24, 64 + minimum_channel_size}, // input size (NHWC)
{96, 3, 3, 64 + minimum_channel_size}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
@@ -363,7 +363,7 @@ struct TestbedConv2dProblemSizes {
// Medium input size, filter size (1x1, 3,x3, 5x5, 7x7), stride (2, 2)
////////////////////////////////////////////////////////////////////////////////////
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 16, 16, 288}, // input size (NHWC)
{1, 13, 16, 288}, // input size (NHWC)
{160, 5, 5, 288}, // filter size (KRSC)
{2, 2, 2, 2}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -371,7 +371,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 55, 55, 256}, // input size (NHWC)
{1, 55, 51, 256}, // input size (NHWC)
{512, 1, 1, 256}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -379,7 +379,7 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 80, 80, 32}, // input size (NHWC)
{1, 71, 80, 32}, // input size (NHWC)
{64, 5, 5, 32}, // filter size (KRSC)
{2, 2, 2, 2}, // padding (pad_h, _, pad_w, _)
{2, 2}, // stride (stride_h, stride_w)
@@ -398,7 +398,7 @@ struct TestbedConv2dProblemSizes {
// Medium input size stride (3, 3), filter (3, 3), non-default padding
////////////////////////////////////////////////////////////////////////////////////
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{1, 27, 27, 256}, // input size (NHWC)
{1, 27, 23, 256}, // input size (NHWC)
{512, 3, 3, 256}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{3, 3}, // stride (stride_h, stride_w)
@@ -464,16 +464,16 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{32, 32, 32, 32}, // input size (NHWC)
{32, 1, 1, 32}, // filter size (KRSC)
{32, 24, 32, 32}, // input size (NHWC)
{32, 1, 2, 32}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
{1, 1} // dilation (dilation_h, dilation_w)
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{4, 3, 3, 128}, // input size (NHWC)
{256, 3, 3, 128}, // filter size (KRSC)
{4, 4, 5, 128}, // input size (NHWC)
{256, 3, 6, 128}, // filter size (KRSC)
{0, 0, 0, 0}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
{1, 1}, // dilation (dilation_h, dilation_w)
@@ -481,8 +481,8 @@ struct TestbedConv2dProblemSizes {
));
conv2d_default_sizes.push_back(cutlass::conv::Conv2dProblemSize(
{4, 1, 1, 256}, // input size (NHWC)
{328, 3, 3, 256}, // filter size (KRSC)
{4, 2, 3, 256}, // input size (NHWC)
{328, 3, 5, 256}, // filter size (KRSC)
{1, 1, 1, 1}, // padding (pad_h, _, pad_w, _)
{1, 1}, // stride (stride_h, stride_w)
{1, 1}, // dilation (dilation_h, dilation_w)