From a08ca63b3645a16edc9efec54682e98795c44efd Mon Sep 17 00:00:00 2001 From: Anca Hamuraru Date: Tue, 20 May 2025 13:14:07 +0000 Subject: [PATCH] Padding not supported for when BDataType is pk_i4_t. Added fix for correct check and removed padding instances. --- .../grid/gridwise_gemm_wmma_cshuffle_v3.hpp | 2 +- .../gpu/gemm_universal.hpp | 22 --------- .../gpu/gemm_universal_wmma.inc | 49 ------------------- .../gpu/gemm_universal/CMakeLists.txt | 28 ----------- ...4_bf16_km_nk_mn_comp_kpadding_instance.cpp | 24 --------- ...bf16_km_nk_mn_comp_mnkpadding_instance.cpp | 25 ---------- ..._bf16_km_nk_mn_comp_mnpadding_instance.cpp | 25 ---------- ...4_bf16_mk_nk_mn_comp_kpadding_instance.cpp | 24 --------- ...bf16_mk_nk_mn_comp_mnkpadding_instance.cpp | 25 ---------- ..._bf16_mk_nk_mn_comp_mnpadding_instance.cpp | 25 ---------- ...i4_f16_km_nk_mn_comp_kpadding_instance.cpp | 24 --------- ..._f16_km_nk_mn_comp_mnkpadding_instance.cpp | 24 --------- ...4_f16_km_nk_mn_comp_mnpadding_instance.cpp | 24 --------- ...i4_f16_mk_nk_mn_comp_kpadding_instance.cpp | 24 --------- ..._f16_mk_nk_mn_comp_mnkpadding_instance.cpp | 24 --------- ...4_f16_mk_nk_mn_comp_mnpadding_instance.cpp | 24 --------- 16 files changed, 1 insertion(+), 392 deletions(-) delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp delete mode 100644 library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp diff --git a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_wmma_cshuffle_v3.hpp b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_wmma_cshuffle_v3.hpp index 4dfa472103..35ecf443d0 100644 --- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_wmma_cshuffle_v3.hpp +++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_wmma_cshuffle_v3.hpp @@ -420,7 +420,7 @@ struct GridwiseGemm_wmma_cshuffle_v3 using GemmSpecialization = tensor_operation::device::GemmSpecialization; - static_assert(!(is_same_v, pk_i4_t> && + static_assert(!(is_same_v, pk_i4_t> && GemmSpec != GemmSpecialization::Default), "pk_i4_t does not support padding"); diff --git a/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal.hpp b/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal.hpp index 49721ed22e..ab5f72f851 100644 --- a/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal.hpp +++ b/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal.hpp @@ -195,24 +195,12 @@ struct DeviceOperationInstanceFactory< { add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_default_instances( op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instances( - op_ptrs); } else if constexpr(is_same_v && is_same_v && is_same_v) { add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_default_instances( op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instances( - op_ptrs); } } #endif @@ -315,21 +303,11 @@ struct DeviceOperationInstanceFactory< is_same_v) { add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_default_instances(op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instances(op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instances( - op_ptrs); } else if constexpr(is_same_v && is_same_v && is_same_v) { add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_default_instances(op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instances(op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instances( - op_ptrs); - add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instances( - op_ptrs); } } #endif diff --git a/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal_wmma.inc b/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal_wmma.inc index e5aaef94e5..75814bd78a 100644 --- a/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal_wmma.inc +++ b/library/include/ck/library/tensor_operation_instance/gpu/gemm_universal_wmma.inc @@ -194,35 +194,10 @@ void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_default_instances std::vector>>& instances); -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances); - void add_device_gemm_wmma_universal_f8_f8_bf16_mk_nk_mn_comp_default_instances( std::vector>>& instances); -void add_device_gemm_wmma_universal_f8_f8_bf16_mk_nk_mn_comp_kpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f8_f8_bf16_mk_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f8_f8_bf16_mk_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances); #endif #if(defined(CK_ENABLE_FP16) && defined(CK_ENABLE_FP8)) void add_device_gemm_wmma_universal_f8_f16_f16_mk_kn_mn_comp_default_instances( @@ -365,35 +340,11 @@ void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_default_instances( std::vector>>& instances); -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances); void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_default_instances( std::vector>>& instances); -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances); -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances); #endif } // namespace instance } // namespace device diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/CMakeLists.txt b/library/src/tensor_operation_instance/gpu/gemm_universal/CMakeLists.txt index 6a9cf9a2bc..c8d56f46be 100644 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/CMakeLists.txt +++ b/library/src/tensor_operation_instance/gpu/gemm_universal/CMakeLists.txt @@ -43,24 +43,10 @@ list(APPEND GEMM_UNIVERSAL_INSTANCES device_gemm_wmma_universal_bf16_bf16_bf16/device_gemm_wmma_universal_bf16_bf16_bf16_km_nk_mn_comp_mnkpadding_instance.cpp device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_default_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_default_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp - device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_default_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_default_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp - device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp device_gemm_wmma_universal_f8_f16_f16/device_gemm_wmma_universal_f8_f16_f16_mk_kn_mn_comp_default_instance.cpp device_gemm_wmma_universal_f8_f16_f16/device_gemm_wmma_universal_f8_f16_f16_mk_kn_mn_comp_kpadding_instance.cpp @@ -199,24 +185,10 @@ set_source_files_properties(device_gemm_wmma_universal_bf16_bf16_bf16/device_gem set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") - set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") - set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") -set_source_files_properties(device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") set_source_files_properties(device_gemm_wmma_universal_f8_f16_f16/device_gemm_wmma_universal_f8_f16_f16_mk_kn_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") set_source_files_properties(device_gemm_wmma_universal_f8_f16_f16/device_gemm_wmma_universal_f8_f16_f16_mk_kn_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1") diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp deleted file mode 100644 index c0a1dd352f..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_kpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp deleted file mode 100644 index 4c87a5c260..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instance.cpp +++ /dev/null @@ -1,25 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, - device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp deleted file mode 100644 index 545363832f..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instance.cpp +++ /dev/null @@ -1,25 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, - device_gemm_wmma_universal_bf16_i4_bf16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp deleted file mode 100644 index 865625c62b..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_kpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp deleted file mode 100644 index 30f521794b..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instance.cpp +++ /dev/null @@ -1,25 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, - device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp deleted file mode 100644 index 1637eb3543..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_bf16_i4_bf16/device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instance.cpp +++ /dev/null @@ -1,25 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, - device_gemm_wmma_universal_bf16_i4_bf16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp deleted file mode 100644 index de7d3dc5bf..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_kpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp deleted file mode 100644 index 34ef320ad1..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp deleted file mode 100644 index 5e29496ef1..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_km_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_km_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp deleted file mode 100644 index 111e35e7fe..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_kpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp deleted file mode 100644 index bdf4090c84..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnkpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck diff --git a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp b/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp deleted file mode 100644 index 4b9d57cbd0..0000000000 --- a/library/src/tensor_operation_instance/gpu/gemm_universal/device_gemm_wmma_universal_f16_i4_f16/device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instance.cpp +++ /dev/null @@ -1,24 +0,0 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved. - -#include "device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn.hpp" -#include "ck/host_utility/device_prop.hpp" - -namespace ck { -namespace tensor_operation { -namespace device { -namespace instance { - -void add_device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_mnpadding_instances( - std::vector>>& - instances) -{ - add_device_operation_instances( - instances, device_gemm_wmma_universal_f16_i4_f16_mk_nk_mn_comp_instances{}); -} - -} // namespace instance -} // namespace device -} // namespace tensor_operation -} // namespace ck