Andreas Karatzas 3775d5fcab [ROCm][CI] Adding test groups for parity with upstream (#47479) ...

Signed-off-by: Andreas Karatzas <akaratza@amd.com>

2026-07-03 19:15:01 +04:00

..

__init__.py

[Perf] fuse kernels in gdn (#37813)

2026-04-02 11:52:18 +00:00

chunk_delta_h.py

[ROCm][CI] Adding test groups for parity with upstream (#47479)

2026-07-03 19:15:01 +04:00

chunk_o.py

[Aiter][ROCm] gdn_linear_attn kernel fusion (#40711)

2026-05-07 23:11:37 -07:00

chunk_scaled_dot_kkt.py

[Bugfix] Fix GDN KKT precision loss on Hopper GPUs by aligning tl.dot operand layout with WGMMA (#42076)

2026-05-09 13:08:46 +00:00

chunk.py

[Aiter][ROCm] gdn_linear_attn kernel fusion (#40711)

2026-05-07 23:11:37 -07:00

cumsum.py

[GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill (#38361)

2026-04-03 13:38:02 +00:00

fused_gdn_prefill_post_conv.py

[Perf] fuse kernels in gdn (#37813)

2026-04-02 11:52:18 +00:00

fused_recurrent.py

[Bugfix] Fix GDN FLA kernel crashes with NULL_BLOCK_ID=0 CUDA graph padding (#39064)

2026-04-11 08:35:19 +00:00

fused_sigmoid_gating.py

[Bugfix] Fix GDN FLA kernel crashes with NULL_BLOCK_ID=0 CUDA graph padding (#39064)

2026-04-11 08:35:19 +00:00

index.py

[Perf] [Bugfix] Fix Triton autotuning in inference for Qwen3.5 (#37338)

2026-03-23 00:37:58 -07:00

kda.py

[Perf][KDA] Fuse gate softplus, chunk-local cumsum, and RCP_LN2 scaling (#43667)

2026-05-28 13:47:08 +00:00

l2norm.py

[Model] Add support for OLMo Hybrid (#32550)

2026-03-05 14:51:06 -05:00

layernorm_guard.py

[Model] Add support for OLMo Hybrid (#32550)

2026-03-05 14:51:06 -05:00

op.py

Update KDA chunk prefill decay to use exp2 semantics (#43195)

2026-05-21 01:16:27 -07:00

solve_tril.py

[GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill (#38361)

2026-04-03 13:38:02 +00:00

utils.py

[Perf][GDN] Align TMA usage with upstream FLA (#38981)

2026-04-05 00:38:02 +08:00

wy_fast.py

[GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill (#38361)

2026-04-03 13:38:02 +00:00