ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-07-12 02:05:50 +00:00
Code Issues Packages Projects Releases Wiki Activity
2,984 Commits 778 Branches 40 Tags
db5cd5ed2de4b9fa10da06b60be49ee8dd292983
Commit Graph

2984 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Qianfeng Zhang
db5cd5ed2d flag mask class interface IsTokenPairInsideMask() as const 2026-07-07 11:13:02 +00:00
Qianfeng Zhang
4948520212 Add leeked including 2026-06-30 14:59:52 +00:00
Qianfeng Zhang
53e69c1404 Use run-time checking (almost_invariant_seqlen) to replace pre-compiling setting (HSTU_SCHED_BATCH_AS_FIRST_GRID_DIM) for switching grid-schedule 2026-06-30 14:06:05 +00:00
Qianfeng Zhang
4667be2bb0 Fix the compiling error in HSTU_SCHED_BATCH_AS_FIRST_GRID_DIM=0 path 2026-06-30 09:10:12 +00:00
Qianfeng Zhang
80d7b8dce7 Adjust the threhold value in shall_use_splitkv() 2026-06-30 09:09:04 +00:00
Qianfeng Zhang
26400def37 Add max_seqlen_kv as host API parameter and adjust the rules for using splitkv for cross-attention 2026-06-29 14:39:57 +00:00
Qianfeng Zhang
7c1c61b710 Fix a bug in calculating max_num_target in run_group_hstu_forward() 2026-06-28 15:07:21 +00:00
Qianfeng Zhang
bdc300591b Removing the class derivation to simplify struct HstuAttentionFwdCommonDropoutKargs 2026-06-24 09:06:01 +00:00
Qianfeng Zhang
db5cb5d19d Update to example_hstu_attention_fwd.cpp 2026-06-22 09:23:23 +00:00
Qianfeng Zhang
e349a39f3d Rename generate_instances.py to generate_fwd_instances.py 2026-06-22 08:02:23 +00:00
Qianfeng Zhang
2694adbd55 Remove the using of kSubQKHeaddim 2026-06-19 05:17:35 +00:00
Qianfeng Zhang
a625506698 Add restriction on the relationship between HstuAttention<xxx>Problem and HstuAttention<xxx>TileSettingClass 2026-06-17 15:59:29 +00:00
Qianfeng Zhang
3eb929334b Add static_assert() in HstuAttentionFwdTileSettingClass 2026-06-14 13:39:22 +00:00
Qianfeng Zhang
c0a011a275 Fix the comments in reference_hstu_attention_fwd.hpp 2026-06-12 14:50:22 +00:00
Qianfeng Zhang
73b5f5ee38 Remove un-used element-wise functions passed through pipelines' operator() interfaces 2026-06-12 08:41:40 +00:00
Qianfeng Zhang
bff80fc39e Rename GetKVBlockGemm to GetPVTBlockGemm 2026-06-12 07:37:18 +00:00
Qianfeng Zhang
de4f776531 Remove the kHasBias==true instances to save building time 2026-06-12 07:37:18 +00:00
Qianfeng Zhang
c7cde5807d Renaming BUILD_HSTU_FOR_GFX95_ONLY to BUILD_HSTU_FOR_GFX95 2026-06-10 15:31:46 +00:00
Qianfeng Zhang
042e35ec9c Update the README.md according to the summary by claude code 2026-06-10 14:42:49 +00:00
Qianfeng Zhang
5204dc7580 Fix the using of num_targets[] in run_group_hstu_attention 2026-06-10 09:00:39 +00:00
Qianfeng Zhang
73965593f9 Renaming in the dispatching codes and generate_instances.py scripts 2026-06-08 15:24:48 +00:00
Qianfeng Zhang
cde6f60611 Split hstu_attention_util.hpp into host_util.hpp and kernel_util.hpp 2026-06-06 06:43:35 +00:00
Qianfeng Zhang
20f97ae684 Remove un-needed includings from some hpp and cpp files 2026-06-06 04:52:38 +00:00
Qianfeng Zhang
797ade1b7f Tiny update in generate_instances.py 2026-06-06 04:52:38 +00:00
Qianfeng Zhang
bd8c2d05fa Fix header file mapping bug in generate_instances.py 2026-06-06 04:52:38 +00:00
Qianfeng Zhang
c71cdbcad7 Some renaming in kernel and pipeline 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
1304e807fb Update and fix for leeked changes and make the scripts be able to test/benchmark kStoreLSE cases 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
798fd3cd8b Enable the kernel dispatching path from is_training & use_softmax to kStoreLSE 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
8b62d651a4 Add instances and kStoreLSE template in dispatcher class to support outputting lse for fwd training 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
c1e3b9be6a Set lse tensor dim strides in example 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
0fdca0e940 Replace template kUseSoftmax/kStoreLSE by boolean parameters in reference fwd codes to save compiling time 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
8f091efc4c Add support for outputing lse in the example and reference hstu attention forward implementation 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
cc184fc202 Add support for preparing lse_dram_window in hstu fwd kernel 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
75a3b5aab0 Kernel use types declared in the problem rather than the pipeline 2026-06-05 15:51:39 +00:00
Qianfeng Zhang
696d4534b2 Tiny simplification with defining the Bias related Kargs 2026-06-03 09:55:04 +00:00
Qianfeng Zhang
eba3c2f635 Add parameters used by storing lse in the fwd and fwd_splitkv_combine kernel to prepare for supporting training 2026-06-03 09:55:04 +00:00
Qianfeng Zhang
5ee8a37cd3 Move num_splits/o_acc_ptr/l_acc_ptr out from HstuAttention<xxx>FwdParams struct 2026-06-03 09:55:04 +00:00
Qianfeng Zhang
36dd77fb16 Add kStoreLSE template parameter to the problems 2026-06-03 09:55:04 +00:00
Qianfeng Zhang
333abddbae Rename the reference interfaces and the files 2026-05-28 08:07:54 +00:00
Qianfeng Zhang
e841981ddd Update to MakeLSEaccDramTileDistribution trying to assign more threads to MThreadPerWarp so that block_tile_reduce_sync() work on less KThreadPerWarp 2026-05-23 07:24:00 +00:00
Qianfeng Zhang
1dbd127d1b Use buffer_view to create lse_acc_dram_naive so that out_of_boundary loading value can be specified (be -inf) 2026-05-23 04:37:52 +00:00
Qianfeng Zhang
86d8d72008 Use partition_index parameter for all get_x_indices_from_distributed_indices() calls 2026-05-22 15:08:30 +00:00
Qianfeng Zhang
65992be728 Update to the cross_attention test/bench scripts 2026-05-22 14:44:14 +00:00
Qianfeng Zhang
b1052e87e1 Add implementation of hstu fwd splitkv for softmax path 2026-05-22 09:54:27 +00:00
Qianfeng Zhang
8a7529177d Fix the calling context for type_context in scale_tile_in_scalar()/scale_tile_in_pack 2026-05-11 08:53:07 +00:00
Qianfeng Zhang
0a32eddc0a Re-format the .hpp/.cpp files using clang-format-18 2026-05-10 13:46:02 +00:00
Qianfeng Zhang
6981f148ee Fix potential bug in kernel host interface BlockSize() 2026-05-08 06:27:19 -04:00
Qianfeng Zhang
250f325c3a More consideration in MakeOaccDramTileDistribution() in splitkv_combine pipeline policy 2026-05-07 06:32:01 -04:00
Qianfeng Zhang
888b6cad86 Use inline-assembly based v_pk_mul_f32 to scale tile pcomp_tile in non-softmax pipeline on gfx950 2026-04-30 14:06:06 +00:00
Qianfeng Zhang
4c583f0574 Add -fno-slp-vectorize option for building hstu kernels on gfx950 2026-04-30 13:37:22 +00:00