3 people 0a6446005d [CPU] Migrate unquantized MoE to the modular-kernel experts structure (#50133) ...

Signed-off-by: jiang1.li <jiang1.li@intel.com>
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Fadi Arafeh <fadi.arafeh@arm.com>

2026-08-03 14:55:04 +08:00

..

modular_kernel_tools

[Log] Update deepgemm log (#45857)

2026-06-17 15:34:22 +00:00

__init__.py

[Kernel] DeepEP dispatch-combine kernel integration (#18434)

2025-06-03 12:30:02 -07:00

conftest.py

[MoE Refactor][Test] FusedMoE layer test (#24675)

2026-04-06 17:17:23 +00:00

parallel_utils.py

[WideEP] Update NCCL to 2.30.7 to enable DeepEPv2 in the vllm/vllm-openai image (#45321)

2026-07-24 13:00:02 -07:00

test_batched_deepgemm.py

[MoE] Remove inplace fused experts mechanism (#43727)

2026-05-27 20:00:19 -07:00

test_batched_moe.py

[Kernel] TD operand loads for batched MoE GEMM (moe_mmk) on XPU (#46340)

2026-07-26 08:50:07 +08:00

test_block_fp8.py

[MoE] Remove inplace fused experts mechanism (#43727)

2026-05-27 20:00:19 -07:00

test_block_int8.py

[CI][Bugfix] Fix test isolation in block_int8/ptpc_fp8 MoE kernel tests (#49609)

2026-07-23 12:26:02 -05:00

test_count_expert_num_tokens.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

test_cpu_fused_moe.py

[CPU] Migrate unquantized MoE to the modular-kernel experts structure (#50133)

2026-08-03 14:55:04 +08:00

test_cpu_int4_moe.py

[CPU][BugFix] Multiple fixes to w4a8_int8 CPU MoE path (#46739)

2026-07-06 05:39:20 +00:00

test_cpu_quant_fused_moe.py

[CPU] Support cpu compressed-tensor w8a8 int8 moe (#42920)

2026-06-29 03:04:05 +00:00

test_cutedsl_moe.py

[CI/Build] Bump flashinfer to v0.6.11.post2 (#41711)

2026-05-16 14:55:12 -07:00

test_cutlass_moe.py

[MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)

2026-06-08 10:42:58 -04:00

test_deepep_deepgemm_moe.py

use split_group for pytorch process group creation (#41980)

2026-06-04 14:36:07 -04:00

test_deepep_moe.py

[Hardware][AMD][Kernel][CI][Bugfix] Fix ROCm DeepEP FP8 max (#50467)

2026-07-30 21:37:36 -05:00

test_deepep_v2_moe.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

test_deepgemm.py

[Kimi K3 Bug] Fix deepgemm support for kimi k3 (#50458)

2026-07-30 20:10:16 -04:00

test_flashinfer_b12x_moe.py

[BugFix] weights processing peak memory reduction for nvfp4 MoE layers (#46276)

2026-07-11 02:05:35 +00:00

test_flashinfer_cutedsl_nvfp4_moe.py

Integrate CuTeDSL MoE for ReLU2 NVFP4 (#49580)

2026-07-28 19:46:38 -07:00

test_flashinfer_moe.py

[MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)

2026-06-08 10:42:58 -04:00

test_flashinfer.py

[MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)

2026-06-08 10:42:58 -04:00

test_flydsl_moe.py

[ROCm][Perf] Enable W4A16 FlyDSL MoE (#44400)

2026-06-14 00:14:39 -07:00

test_fused_topk.py

test: add nan/inf clamp regression test for fused_topk_bias (#40553)

2026-04-22 00:46:53 +00:00

test_gemma4router.py

[FEAT] [Perf] [Gemma4] Fused Gemma4 Routing Function Triton (#39083)

2026-04-19 09:57:39 +00:00

test_gpt_oss_triton_kernels.py

[MoE] Move GPT OSS Triton kernel experts into fused_moe/experts/ (#39007)

2026-04-14 19:27:39 +00:00

test_grouped_topk.py

[Model] Add Kimi K3 support: model files and kernels [1/N] (#50089)

2026-07-29 14:10:58 +08:00

test_marlin_vs_trtllm_mxint4.py

[Kernel][MoE] Allow FlashInfer MXINT4 MoE for gated SiLU (#46518)

2026-06-24 18:32:50 -05:00

test_modular_kernel_combinations.py

[MoE Refactor] Remove supports_expert_map (#43108)

2026-05-29 17:26:56 -04:00

test_modular_oai_triton_moe.py

Enable gfx1250 ROCm architecture (#46516)

2026-07-31 01:41:12 +00:00

test_moe_align_block_size.py

[Model] Add Kimi K3 support: model files and kernels [1/N] (#50089)

2026-07-29 14:10:58 +08:00

test_moe_kernel_oracle.py

[MoE] [MoE Refactor] Add moe kernel oracle abc 37753 (#43461)

2026-06-25 09:34:03 -04:00

test_moe_layer.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

test_moe_permute_unpermute.py

[Perf] Optimize moe permute by pre-allocate buffer, 9~14% kernel performance improvement (#43014)

2026-05-28 06:18:26 -07:00

test_moe_weight_loading_padded.py

[Bugfix] Don't transpose fused MoE quantization scales in RoutedExperts.load_weights (#50137)

2026-07-31 08:46:52 -07:00

test_moe.py

[CPU] Migrate unquantized MoE to the modular-kernel experts structure (#50133)

2026-08-03 14:55:04 +08:00

test_mxfp4_moe.py

Fix the E8M0 scale computation in the MXFP4 (W4A4) MOE CUTLASS kernel (#43557)

2026-06-15 06:04:54 -07:00

test_mxfp8_aiter_backend_selection.py

[CI] Fix MXFP8 MOE backend selection tests on gfx942 (#50222)

2026-07-29 17:17:11 +08:00

test_nvfp4_moe.py

[MoE] Remove inplace fused experts mechanism (#43727)

2026-05-27 20:00:19 -07:00

test_ocp_mx_moe.py

[AMD] Revert Mxfp4MoeBackend.TRITON_UNFUSED fallback (#46491)

2026-07-27 22:31:23 -05:00

test_profile_modular_kernel.py

[Bugfix] Fix broken profile_modular_kernel.py (#43300)

2026-06-11 12:17:23 +01:00

test_rocm_aiter_topk.py

[MoE] Move various experts classes to fused_moe/experts/ (#41979)

2026-05-11 07:54:33 +08:00

test_routed_experts_capture_monolithic.py

[Bugfix][Test] Fix monolithic routing replay test buffer capacity (#50640)

2026-07-31 19:37:39 -07:00

test_routing_simulator.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

test_routing.py

Enable gfx1250 ROCm architecture (#46516)

2026-07-31 01:41:12 +00:00

test_shared_fused_moe_routed_transform.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

test_silu_mul_fp8_quant_deep_gemm.py

[UX] Integrate DeepGEMM into vLLM wheel via CMake (#37980)

2026-04-08 18:56:32 -07:00

test_silu_mul_per_token_group_quant_fp8_colmajor.py

[Bugfix] Add swiglu limits to deepgemm fp8 methods (#41986)

2026-05-14 15:43:13 -07:00

test_topk_softplus_sqrt.py

skip cudagraph/DP padding in topk (#48979)

2026-07-21 15:20:30 -07:00

test_triton_moe_no_act_mul.py

[MoE] Move various experts classes to fused_moe/experts/ (#41979)

2026-05-11 07:54:33 +08:00

test_triton_moe_ptpc_fp8.py

[CI][Bugfix] Fix test isolation in block_int8/ptpc_fp8 MoE kernel tests (#49609)

2026-07-23 12:26:02 -05:00

test_trtllm_bf16_moe.py

[Perf] Integrate TRTLLM BF16 MoE Modular Kernel (#45182)

2026-07-09 01:36:14 +04:00

test_trtllm_nvfp4_moe.py

[MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)

2026-06-08 10:42:58 -04:00

test_unquantized_backend_selection.py

[CPU] Migrate unquantized MoE to the modular-kernel experts structure (#50133)

2026-08-03 14:55:04 +08:00

test_zero_expert_moe.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

utils.py

Enable B12x backend for non-gated MoEs (like Nemotron) (#43328)

2026-07-06 12:40:07 -07:00