Skip to content

GitLab

"tests/pipelines/vscode:/vscode.git/clone" did not exist on "e4b8e7928b2c1972b37af67c64ccc67e42578f8c"
Commit 4976eb0c authored by wangshaojie6's avatar wangshaojie6
Browse files

add test and instance for MNK padding

parent 812521df
Hide whitespace changes
Inline Side-by-side
Showing with 16 additions and 3 deletions
library/src/tensor_operation_instance/gpu/batched_gemm_masking_scale_softmax_gemm_permute/device_batched_gemm_masking_scale_softmax_gemm_permute_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instance.cpp
View file @ 4976eb0c
...@@ -31,7 +31,7 @@ using Scale = ck::tensor_operation::element_wise::Scale; ...@@ -31,7 +31,7 @@ using Scale = ck::tensor_operation::element_wise::Scale;
static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default; static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
static constexpr auto GemmPadded = static constexpr auto GemmPadded =
ck::tensor_operation::device::GemmSpecialization::MNOPadding; // Padding K is currently flawed ck::tensor_operation::device::GemmSpecialization::MNKOPadding; // Padding K is currently flawed
// c[g, m, n] = a[g, m, k] * b[g, n, k] // c[g, m, n] = a[g, m, k] * b[g, n, k]
using device_batched_gemm_masking_scale_softmax_gemm_permute_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instances = using device_batched_gemm_masking_scale_softmax_gemm_permute_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instances =
...@@ -44,9 +44,9 @@ using device_batched_gemm_masking_scale_softmax_gemm_permute_xdl_cshuffle_f16_f1 ...@@ -44,9 +44,9 @@ using device_batched_gemm_masking_scale_softmax_gemm_permute_xdl_cshuffle_f16_f1
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 64, 32, 8, 8, 2, 32, 32, 2, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 64, 32, 8, 8, 2, 32, 32, 2, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 128, 32, 8, 8, 2, 32, 32, 2, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 256, 128, 32, 128, 32, 8, 8, 2, 32, 32, 2, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 64, 32, 8, 8, 2, 32, 32, 1, 8, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 64, 32, 8, 8, 2, 32, 32, 1, 8, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 128, 32, 8, 8, 2, 32, 32, 1, 8, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, //DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 256, 32, 128, 32, 8, 8, 2, 32, 32, 1, 8, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 64, 64, 32, 8, 8, 2, 32, 32, 1, 4, 2, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 64, 64, 32, 8, 8, 2, 32, 32, 1, 4, 2, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 64, 32, 8, 8, 2, 32, 32, 1, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, //DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 64, 32, 8, 8, 2, 32, 32, 1, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<16, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 64, 128, 32, 8, 8, 2, 32, 32, 1, 4, 4, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 64, 128, 32, 8, 8, 2, 32, 32, 1, 4, 4, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, false, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 128, 32, 8, 8, 2, 32, 32, 1, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 128, 128, 32, 128, 32, 8, 8, 2, 32, 32, 1, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 2, S<1, 32, 1, 8>, 8, true>,
DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 64, 256, 32, 128, 32, 8, 8, 2, 16, 16, 1, 16, 8, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 8, S<1, 16, 1,16>, 8, true>, DeviceBatchedGemmSoftmaxGemmPermute_Xdl_CShuffle< Row, Col, Row, CPermuteNumDims_G_M_O, F16, F16, F16, F16, F32, F16, PassThrough, PassThrough, Scale, PassThrough, PassThrough, GemmDefault, 1, 256, 64, 256, 32, 128, 32, 8, 8, 2, 16, 16, 1, 16, 8, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S< 8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 2, false, 1, 8, S<1, 16, 1,16>, 8, true>,
......
test/batched_gemm_masking_scale_softmax_gemm_permute/test_batched_gemm_masking_scale_softmax_gemm_permute_fp16.cpp
View file @ 4976eb0c
...@@ -91,6 +91,19 @@ TYPED_TEST(TestBatchedGemmMaskingScaleSoftmaxGemmPermuteFP16, Test_FP16_OddO) ...@@ -91,6 +91,19 @@ TYPED_TEST(TestBatchedGemmMaskingScaleSoftmaxGemmPermuteFP16, Test_FP16_OddO)
this->Run(); this->Run();
} }
TYPED_TEST(TestBatchedGemmMaskingScaleSoftmaxGemmPermuteFP16, DISABLED_Bench_FP16_IrregularK)
{
this->lengths_ = std::vector<std::vector<int>>{{256, 256, 160, 160, 1, 16},
{256, 64, 160, 64, 1, 16},
{1024, 1024, 80, 80, 1, 16},
{1024, 64, 80, 64, 1, 16},
{4096, 4096, 40, 40, 1, 16},
{4096, 64, 40, 64, 1, 16}};
this->bench_ = true;
this->verify_ = false;
this->Run();
}
TYPED_TEST(TestBatchedGemmMaskingScaleSoftmaxGemmPermuteFP16, DISABLED_Bench_FP16) TYPED_TEST(TestBatchedGemmMaskingScaleSoftmaxGemmPermuteFP16, DISABLED_Bench_FP16)
{ {
this->lengths_ = std::vector<std::vector<int>>{ this->lengths_ = std::vector<std::vector<int>>{
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment