mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-05 14:11:29 +00:00
[CK_TILE] moe sorting ex kernel to support expert > 128 (#1840)
* moe sorting ex * fix bug for race condition * fix bug and optimze large expert * fix * optimize with sub_token_oneshot * support skip empty tokens for expert sorting * update moe_sorting * tidy code
This commit is contained in:
@@ -7,6 +7,7 @@
|
||||
#include "ck_tile/ops/fused_moe/kernel/fused_moegemm_shape.hpp"
|
||||
#include "ck_tile/ops/fused_moe/kernel/fused_moegemm_tile_partitioner.hpp"
|
||||
#include "ck_tile/ops/fused_moe/kernel/moe_sorting_kernel.hpp"
|
||||
#include "ck_tile/ops/fused_moe/kernel/moe_sorting_problem.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/fused_moegemm_pipeline_flatmm_ex.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/fused_moegemm_pipeline_flatmm_policy.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/fused_moegemm_pipeline_flatmm_uk.hpp"
|
||||
@@ -14,7 +15,6 @@
|
||||
#include "ck_tile/ops/fused_moe/pipeline/fused_moegemm_traits.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/moe_sorting_pipeline.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/moe_sorting_policy.hpp"
|
||||
#include "ck_tile/ops/fused_moe/pipeline/moe_sorting_problem.hpp"
|
||||
#include "ck_tile/ops/common/generic_2d_block_shape.hpp"
|
||||
#include "ck_tile/ops/common/tensor_layout.hpp"
|
||||
#include "ck_tile/ops/common/utils.hpp"
|
||||
|
||||
Reference in New Issue
Block a user