merge with the develop support the fp8 with computev4

9ba504b6 · ThomasNing · e3402c93 · f49de496 · 9ba504b6 · 9ba504b6
Commit 9ba504b6 authored Feb 07, 2025 by ThomasNing
20 changed files
--- a/include/ck_tile/core/numeric/pk_int4.hpp
+++ b/include/ck_tile/core/numeric/pk_int4.hpp
--- a/include/ck_tile/core/numeric/vector_type.hpp
+++ b/include/ck_tile/core/numeric/vector_type.hpp
--- a/include/ck_tile/core/utility/transpose_vectors.hpp
+++ b/include/ck_tile/core/utility/transpose_vectors.hpp
--- a/include/ck_tile/host/check_err.hpp
+++ b/include/ck_tile/host/check_err.hpp
--- a/include/ck_tile/host/reference/reference_gemm.hpp
+++ b/include/ck_tile/host/reference/reference_gemm.hpp
--- a/include/ck_tile/ops/epilogue/cshuffle_epilogue.hpp
+++ b/include/ck_tile/ops/epilogue/cshuffle_epilogue.hpp
--- a/include/ck_tile/ops/flatmm/block/uk/flatmm_sn_uk_gfx9_32x128x512_1x4x1_16x16x16.inc
+++ b/include/ck_tile/ops/flatmm/block/uk/flatmm_sn_uk_gfx9_32x128x512_1x4x1_16x16x16.inc
--- a/include/ck_tile/ops/flatmm/block/uk/flatmm_sn_uk_gfx9_32x128x512_1x4x1_16x16x16_itl.inc
+++ b/include/ck_tile/ops/flatmm/block/uk/flatmm_sn_uk_gfx9_32x128x512_1x4x1_16x16x16_itl.inc
--- a/include/ck_tile/ops/flatmm/block/uk/flatmm_uk_gfx9_32x512x128_1x1x1_16x16x16.inc
+++ b/include/ck_tile/ops/flatmm/block/uk/flatmm_uk_gfx9_32x512x128_1x1x1_16x16x16.inc
--- a/include/ck_tile/ops/gemm/block/block_universal_gemm_as_bs_cr.hpp
+++ b/include/ck_tile/ops/gemm/block/block_universal_gemm_as_bs_cr.hpp
--- a/include/ck_tile/ops/gemm/kernel/gemm_kernel.hpp
+++ b/include/ck_tile/ops/gemm/kernel/gemm_kernel.hpp
--- a/include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_comp_v3.hpp
+++ b/include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_comp_v3.hpp
--- a/include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_mem.hpp
+++ b/include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_mem.hpp
--- a/include/ck_tile/ops/gemm/pipeline/gemm_universal_pipeline_ag_bg_cr_policy.hpp
+++ b/include/ck_tile/ops/gemm/pipeline/gemm_universal_pipeline_ag_bg_cr_policy.hpp
--- a/library/include/ck/library/reference_tensor_operation/cpu/reference_mx_gemm.hpp
+++ b/library/include/ck/library/reference_tensor_operation/cpu/reference_mx_gemm.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_two_stage_xdl_instance.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_two_stage_xdl_instance.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_xdl_comp_instance.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_xdl_comp_instance.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_xdl_merged_groups_instance.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_xdl_merged_groups_instance.hpp
--- a/library/src/tensor_operation_instance/gpu/CMakeLists.txt
+++ b/library/src/tensor_operation_instance/gpu/CMakeLists.txt
--- a/library/src/tensor_operation_instance/gpu/batched_gemm/device_batched_gemm_xdl_f16_f16_f16_gkm_gkn_gmn_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/batched_gemm/device_batched_gemm_xdl_f16_f16_f16_gkm_gkn_gmn_instance.cpp