ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-07-14 11:07:44 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
dd178ef79d70f415b126791ba39fe4019ea504f9
composable_kernel/include/ck/tensor_operation/gpu/grid
History
mtgu0705 04ff8111d8 Revert "fix nswizzle = true" 
This reverts commit 24d8024f0e.
2025-02-19 16:42:06 +08:00
..
batchnorm_multiblock
…
gemm_layernorm
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
normalization
…
block_to_ctile_map.hpp
Codegen hipRTC compilation (#1579)
2025-01-31 09:48:39 -08:00
gridwise_2d_multiple_reduction_multiblock.hpp
…
gridwise_2d_multiple_reduction_threadwise.hpp
…
gridwise_2d_reduction_multiblock.hpp
…
gridwise_2d_reduction_threadwise_multi_d.hpp
…
gridwise_2d_reduction_threadwise.hpp
…
gridwise_batched_gemm_gemm_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_batched_gemm_multiple_d_gemm_multiple_d_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_batched_gemm_multiple_d_softmax_gemm_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_batched_gemm_softmax_gemm_wmma_cshuffle.hpp
…
gridwise_batched_gemm_softmax_gemm_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_batchnorm_backward_blockwise_welford.hpp
…
gridwise_batchnorm_forward_blockwise_welford.hpp
…
gridwise_elementwise_1d_scale.hpp
…
gridwise_elementwise_2d.hpp
…
gridwise_elementwise_layernorm_welford_variance.hpp
…
gridwise_fpAintB_gemm_wmma.hpp
…
gridwise_gemm_bias_add_reduce_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_dl_multiple_d.hpp
…
gridwise_gemm_dl_v1r3.hpp
…
gridwise_gemm_dpp.hpp
…
gridwise_gemm_multiple_abd_xdl_cshuffle.hpp
Codegen hipRTC compilation (#1579)
2025-01-31 09:48:39 -08:00
gridwise_gemm_multiple_d_multiple_r_xdl_cshuffle.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_multiple_d_wmma_cshuffle.hpp
…
gridwise_gemm_multiple_d_xdl_cshuffle_lds_direct_load.hpp
Change flag to CK_GFX90A_DENORM_WORKAROUND (#1817)
2025-01-28 09:58:39 -05:00
gridwise_gemm_multiple_d_xdl_cshuffle.hpp
Codegen hipRTC compilation (#1579)
2025-01-31 09:48:39 -08:00
gridwise_gemm_multiple_d_xdl_splitk_cshuffle.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_pipeline_selector.hpp
Codegen hipRTC compilation (#1579)
2025-01-31 09:48:39 -08:00
gridwise_gemm_pipeline_v1.hpp
…
gridwise_gemm_pipeline_v2.hpp
…
gridwise_gemm_pipeline_v3.hpp
…
gridwise_gemm_pipeline_v4_direct_load.hpp
…
gridwise_gemm_reduce_xdl_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_split_k_multiple_d_xdl_cshuffle_v2.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_split_k_multiple_d_xdl_cshuffle.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_waveletmodel.hpp
…
gridwise_gemm_wmma.hpp
…
gridwise_gemm_xdl_cshuffle_bwd_weight_v3.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdl_cshuffle_streamk_v3.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdl_cshuffle_v1.hpp
…
gridwise_gemm_xdl_cshuffle_v2.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdl_cshuffle_v3_b_preshuffle.hpp
move b thread dequant copy to blockwise.
2025-02-18 13:25:06 +08:00
gridwise_gemm_xdl_cshuffle_v3_b_scale.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdl_cshuffle_v3_multi_abd.hpp
…
gridwise_gemm_xdl_cshuffle_v3_multi_d_ab_scale.hpp
OCP FP8 support for gfx12. (#1710)
2024-12-03 08:42:55 -08:00
gridwise_gemm_xdl_cshuffle_v3_multi_d_b_preshuffle.hpp
chage ktile
2025-02-17 14:26:43 +00:00
gridwise_gemm_xdl_cshuffle_v3_multi_d.hpp
revert blockwisegemm modification
2025-02-07 02:50:26 +00:00
gridwise_gemm_xdl_cshuffle_v3.hpp
merge develop
2025-02-15 03:23:00 +00:00
gridwise_gemm_xdl_layernorm_cshuffle_v1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdl_waveletmodel_cshuffle.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdlops_bwd_weight.hpp
Change flag to CK_GFX90A_DENORM_WORKAROUND (#1817)
2025-01-28 09:58:39 -05:00
gridwise_gemm_xdlops_skip_b_lds_v1.hpp
…
gridwise_gemm_xdlops_splitk_lds_direct_load.hpp
…
gridwise_gemm_xdlops_streamk.hpp
…
gridwise_gemm_xdlops_v2r3.hpp
Change flag to CK_GFX90A_DENORM_WORKAROUND (#1817)
2025-01-28 09:58:39 -05:00
gridwise_gemm_xdlops_v2r4.hpp
…
gridwise_gemm_xdlops_v2r4r2.hpp
…
gridwise_gemm_xdlops_v3r1.hpp
Fix KPack and enable existing instances on gfx950 (#1871)
2025-02-12 09:46:38 -08:00
gridwise_gemm_xdlops_v3r2.hpp
…
gridwise_gemm_xdlops_v3r3.hpp
…
gridwise_moe_gemm_gather.hpp
fix bug in moe_gemm1.cpp, now function pass.
2025-02-18 13:32:46 +08:00
gridwise_moe_gemm_scatter.hpp
Added moe_pk_i4_gemm2, function pass.
2025-02-18 13:34:41 +08:00
gridwise_moe_gemm.hpp
Revert "fix nswizzle = true"
2025-02-19 16:42:06 +08:00
gridwise_permute.hpp
…
gridwise_put_element_1d.hpp
…
gridwise_set_buffer_value.hpp
…
gridwise_set_multiple_buffer_value.hpp
…
gridwise_softmax.hpp
…
gridwise_sparse_embeddings_forward_layernorm.hpp
…
gridwise_tensor_rearrange.hpp
…