Thomas Ning 64a2fda713 Revert "Fix default epilogue (#2358)" (#2364) ...

This reverts commit cd606f72c1.

2025-06-17 22:43:05 -07:00

..

add_rmsnorm2d_rdquant

[A8W8 GEMM] Optimized weight-preshuffled implementation & add quantization datatype for CK TILE rms_norm (#1862)

2025-02-20 14:00:27 -08:00

batched_transpose

Vectorized Transpose for Batched Transpose CK Tile Operator (#2131)

2025-05-12 00:41:45 -07:00

common

chore: unset executable permission (#2303)

2025-06-10 09:13:59 -07:00

elementwise

[CK_TILE] Multiple-D GEMM example (#2219)

2025-06-13 19:39:11 +02:00

epilogue

Revert "Fix default epilogue (#2358)" (#2364)

2025-06-17 22:43:05 -07:00

flatmm

Do not use warpSize as compile time constant as it is removed (#2320)

2025-06-17 11:54:30 -07:00

fmha

Do not use warpSize as compile time constant as it is removed (#2320)

2025-06-17 11:54:30 -07:00

fused_moe

[CK_TILE] moe_sorting support "local_tokens" feature for EP case (#2335)

2025-06-18 10:49:43 +08:00

gemm

[CK_TILE] Support multi-config in tile_example_gemm_universal (#2240)

2025-06-17 17:27:46 -07:00

image_to_column

Do not use warpSize as compile time constant as it is removed (#2320)

2025-06-17 11:54:30 -07:00

layernorm2d

[CK_TILE] Improve RMS/Layer Normalization 2 Pass Pipeline Performance (#1861)

2025-03-25 20:09:45 +08:00

norm_reduce

Do not use warpSize as compile time constant as it is removed (#2320)

2025-06-17 11:54:30 -07:00

permute

…

reduce/block

Do not use warpSize as compile time constant as it is removed (#2320)

2025-06-17 11:54:30 -07:00

rmsnorm2d

[CK_TILE] Improve RMS/Layer Normalization 2 Pass Pipeline Performance (#1861)

2025-03-25 20:09:45 +08:00

smoothquant

…

softmax/block

…

topk/block

…

topk_softmax

…

add_rmsnorm2d_rdquant.hpp

…

batched_transpose.hpp

…

common.hpp

…

elementwise.hpp

…

epilogue.hpp

Solve the Static Encoding Pattern compile error when the tile size is too small (#2079)

2025-04-13 20:09:30 -07:00

flatmm.hpp

[flatmm] implement basic fp16 flatmm (#2089)

2025-04-16 16:51:17 +08:00

fmha.hpp

[CK_TILE] Add logits soft-capping & customization support to the FMHA forward kernel/pipelines (#2163)

2025-05-13 12:19:25 +08:00

fused_moe.hpp

[CK_TILE] moe sorting ex kernel to support expert > 128 (#1840)

2025-02-11 17:49:17 +08:00

gemm.hpp

Fix the CK Tile related operators (#2356)

2025-06-16 17:38:52 -07:00

image_to_column.hpp

…

layernorm2d.hpp

…

norm_reduce.hpp

…

permute.hpp

…

reduce.hpp

…

rmsnorm2d.hpp

…

smoothquant.hpp

…

softmax.hpp

…

topk_softmax.hpp

…

topk.hpp

…