3 people 63b9931b73 [ROCm][DSv4.1] Paged MXFP4 sparse indexer on aiter's MQA-logits kernel (#58671) ...

Signed-off-by: Mehmet Cagri Kaymak <mehmet.kaymak@amd.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Shanshan Shen <467638484@qq.com>

2026-09-29 14:37:57 +08:00

..

conftest.py

[Chore] Clean up pytorch helper functions in vllm.utils (#26908)

2025-10-18 09:48:22 -07:00

test_amx_mla.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_attention_selector.py

[Refactor] Remove dead or duplicate tests (#58446)

2026-09-24 10:29:56 -04:00

test_attention.py

[Refactor] Remove dead tests utils (#58803)

2026-09-26 09:16:53 -04:00

test_cache.py

[Bugfix][SM120][MLA] Support NoPE sparse MLA (GLM-5.3-Flash) on the FlashInfer SM120 backend (#55277)

2026-09-23 02:03:20 -07:00

test_cascade_flash_attn.py

[Platform] Deprecate seed_everything (#31659)

2026-01-04 18:34:04 -08:00

test_cpu_attn.py

[Attention][CPU] Run Zen CPU encoder attention on zentorch SDPA (#54508)

2026-09-24 02:05:44 -07:00

test_cutlass_mla_decode.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_deepgemm_attention.py

[Attention][DSA] Take the native decode path for MTP=3 on SM90 (#52164)

2026-08-14 10:09:51 +00:00

test_flash_attn.py

[SM100] Hdim 256 optimized (#52980)

2026-08-25 20:28:40 -07:00

test_flashinfer_mla_decode.py

[Perf][Attention] Reduce GLM sparse MLA preparation overhead (#57458)

2026-09-21 17:24:25 -07:00

test_flashinfer_trtllm_attention.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_flashinfer.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_flashmla_sparse.py

[Perf][DSv4] Fuse inverse RoPE + FP8 quant into FlashInfer sparse MLA (#58621)

2026-09-24 23:26:50 -07:00

test_flashmla.py

[ROCM][CI] Attention test speedup (#52763)

2026-08-18 16:40:30 -05:00

test_kimi_k3_mla_fused_epilogue.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_kimi_k3_mla_key_concat_kv_cache.py

[Bugfix] Fix SM100 fp8_ds_mla cache scales (#49435)

2026-09-19 17:15:07 -07:00

test_lightning_attn.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

test_merge_attn_states.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_mha_attn.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

test_minimax_m3_fp8_triton_indexer.py

[Minimax M3] Enable fp8 indexer cache on triton indexer for non-SM100 architectures. (#59081)

2026-09-28 19:21:13 -04:00

test_minimax_m3_msa_cutlass_sparse_decode.py

[Bugfix][MiniMax-M3] Keep FP8 query allocation stable across CUDA graph replay (#51203)

2026-08-20 10:45:28 -07:00

test_minimax_m3.py

[AMD][Minimax-M3][perf] Enable packed LBHNC AITER QK-norm fusion for MiniMax-M3 on ROCm (#54535)

2026-09-21 16:38:30 +08:00

test_mixed_causal_attn.py

[Model] Add DiffusionGemma Support (#45163)

2026-06-11 22:17:35 -07:00

test_mla_cross_layer_kernel_equivalence.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_mla_decode_cpu.py

[Chore]:Extract math and argparse utilities to separate modules (#27188)

2025-10-26 04:03:32 -07:00

test_pack_unpack_triton.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

test_prefix_prefill.py

[Kernel] Make prefix-prefill tiling independent of the KV page size (#54194)

2026-09-01 09:19:55 +08:00

test_rocm_aiter_candidate_mask.py

[ROCm][DSV4.1][Perf] Stride the DSA decode candidate mask over the live context (#56628)

2026-09-14 20:47:32 +00:00

test_rocm_aiter_fa.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_rocm_aiter_mla_causal_verify_mask.py

[ROCm][MLA][DCP] Support causal multi-token verification (#51705)

2026-08-31 10:42:28 -07:00

test_rocm_aiter_mla_decode_metadata.py

[Bugfix][ROCm][CI] Give the AITER MLA decode metadata stub its MLA dims (#52139)

2026-08-13 13:35:00 -05:00

test_rocm_aiter_mla_decode.py

[ROCm][CI] Add MLA decode accuracy and determinism tests (#50480)

2026-08-05 20:35:56 -05:00

test_rocm_aiter_mla_fp8_prefill.py

[ROCm][K3] Extend FP8 asm MLA prefill to non-divisor small head counts (#51040)

2026-08-26 17:47:16 +00:00

test_rocm_aiter_mla_fp8_support.py

[ROCm][Spec Decode] Add Aiter MLA decode support non-causal draft block (#55966)

2026-09-15 12:18:09 -05:00

test_rocm_aiter_mla_head_padding.py

[ROCm][MLA][DCP] Support causal multi-token verification (#51705)

2026-08-31 10:42:28 -07:00

test_rocm_aiter_mla_op_registration.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

test_rocm_aiter_mla_sink.py

[ROCm][CI] Add attention-sink support to ROCm AITER sparse MLA (#54404)

2026-09-07 09:26:47 +00:00

test_rocm_aiter_mla_sparse_metadata_sync.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_rocm_aiter_unified_attn.py

Fix ROCm AITER FP8 KV test tolerances. (#53590)

2026-09-09 12:22:25 -05:00

test_rocm_paged_mxfp4_indexer.py

[ROCm][DSv4.1] Paged MXFP4 sparse indexer on aiter's MQA-logits kernel (#58671)

2026-09-29 14:37:57 +08:00

test_rocm_triton_attn_dsv4.py

[Bugfix][ROCm] Drop -1 sentinels when building the ragged sparse-MLA indices (#58058)

2026-09-28 18:40:11 -05:00

test_triton_decode_attention.py

[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

2026-09-16 10:57:24 -07:00

test_triton_mla_causal_verify_flatten.py

[Bugfix][MLA] TritonMLA: fix illegal memory access on causal multi-token decode (#51065)

2026-09-18 03:49:53 -05:00

test_triton_prefill_attention.py

[Perf][ROCm][Attention] Narrow the Triton prefill-attention KV tile on RDNA3/RDNA4 (#58225)

2026-09-23 08:39:27 +00:00

test_triton_unified_attention_diffkv.py

[CI][ROCm] Mirror remaining portable suites without duplicate coverage (#57877)

2026-09-21 04:16:26 -05:00

test_triton_unified_attention.py

[Bugfix][Attention] Avoid NaN in the Triton softcap for large attention logits (#56579)

2026-09-22 12:11:05 -04:00

test_trtllm_kvfp8_dequant.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

test_use_trtllm_attention.py

[Attention] Add FlashInfer XQA decode support on SM12x (#49718)

2026-08-11 16:32:49 -07:00

test_xpu_mla_sparse.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00