carlushuang 807501ac3d [CK_TILE] optimize moe sorting kernel, boost large context case up to 20x (#2153) ...

* combine 2-3 as single stage

* support zeroing

* improve long tokens

* update specialization

* b16 ws

* 8bit topk optimize

* update 15 example

[ROCm/composable_kernel commit: 4e9b76f88c]

2025-05-06 17:32:07 +08:00

..

amd_buffer_addressing_builtins.hpp

Reapply "[CK_TILE] support hdim=192/128 pair for deepseekv3 (#1961)" … (#1971)

2025-03-13 11:41:39 +08:00

amd_buffer_addressing.hpp

Replace inline assembly with builtins in FHMA (#2067)

2025-04-10 09:48:37 +02:00

arch.hpp

[CK_TILE] optimize moe sorting kernel, boost large context case up to 20x (#2153)

2025-05-06 17:32:07 +08:00

generic_memory_space_atomic.hpp

Addressing (Post Merge) code review comments for PR 1845 (#1883)

2025-03-06 11:40:30 -08:00

utility.hpp

[CK_TILE] fused-moe first version (#1634)

2024-11-26 11:14:56 +08:00

workgroup_barrier.hpp

[CK_TILE] optimize moe sorting kernel, boost large context case up to 20x (#2153)

2025-05-06 17:32:07 +08:00