Andrii SkliarandAndrii Skliar b1384f5ec6 Enable B12x backend for non-gated MoEs (like Nemotron) (#43328) ...

Signed-off-by: Andrii Skliar <askliar@nvidia.com>
Co-authored-by: Andrii Skliar <askliar@nvidia.com>

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

..

modular_kernel_tools

[Log] Update deepgemm log (#45857)

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

__init__.py

…

conftest.py

…

parallel_utils.py

[WideEP] Integrate DeepEP v2 (#41183)

2026-06-08 18:07:29 -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

…

test_block_fp8.py

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

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

test_block_int8.py

…

test_count_expert_num_tokens.py

…

test_cpu_fused_moe.py

[CPU][Perf] Accelerate unquantized MoE for AArch64 (#46353)

2026-06-24 14:14:35 +00:00

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

[ROCm][CI TG] refactor and fix deepep_moe test group (#46758)

2026-06-27 10:45:23 +08:00

test_deepep_v2_moe.py

[WideEP] Integrate DeepEP v2 (#41183)

2026-06-08 18:07:29 -07:00

test_deepgemm.py

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

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

test_flashinfer_b12x_moe.py

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

2026-07-06 12:40:07 -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_gemma4router.py

…

test_gpt_oss_triton_kernels.py

…

test_grouped_topk.py

…

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

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

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

test_moe_align_block_size.py

…

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

[EPLB] Mask padding in EPLB load recording (#38128)

2026-06-28 19:43:58 -07:00

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] FusedMoE: coerce shape-(1,) per-tensor scales to 0-D scalar … (#43362)

2026-06-22 13:26:53 -07:00

test_moe.py

[Platform] Replace torch.cuda.mem_get_info with torch.accelerator.get_memory_info (#44825)

2026-06-30 14:39:52 +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

[ROCm][Perf] Use flydsl moe with Minimax-M3 mxfp8 weights on gfx950 and implemented moe-backend selection (#46184)

2026-06-27 10:22:57 +00:00

test_nvfp4_moe.py

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

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

test_ocp_mx_moe.py

[ROCm][CI] Fix Kernels and Kernels attention test failures (#47519)

2026-07-04 15:59:51 -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

…

test_routing_simulator.py

…

test_routing.py

[EPLB] Mask padding in EPLB load recording (#38128)

2026-06-28 19:43:58 -07:00

test_shared_fused_moe_routed_transform.py

…

test_silu_mul_fp8_quant_deep_gemm.py

…

test_silu_mul_per_token_group_quant_fp8_colmajor.py

…

test_topk_softplus_sqrt.py

…

test_triton_moe_no_act_mul.py

…

test_triton_moe_ptpc_fp8.py

…

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

Deprecations for v0.23 and v0.24 (#44992)

2026-06-11 14:35:38 +00:00

test_zero_expert_moe.py

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

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

utils.py

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

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