Flashinfer_CUTLASS_MOE fuses quantization for TP (#27223)

Signed-off-by: Shu Wang. <shuw@nvidia.com>

vllm/model_executor/layers/fused_moe/flashinfer_cutlass_moe.py

@@ -56,6 +56,7 @@ class FlashInferExperts(mk.FusedMoEPermuteExpertsUnpermute):
         ep_size: int = 1,
         tp_rank: int = 0,
         tp_size: int = 1,
+        use_dp: bool = False,
     ):
         super().__init__(quant_config)
         assert quant_config.quant_dtype in ("nvfp4", torch.float8_e4m3fn, None), (
@@ -67,6 +68,7 @@ class FlashInferExperts(mk.FusedMoEPermuteExpertsUnpermute):
         self.tp_rank = tp_rank
         self.tp_size = tp_size
         self.out_dtype = out_dtype
+        self.use_dp = use_dp
 
     @property
     def activation_formats(
@@ -117,7 +119,8 @@ class FlashInferExperts(mk.FusedMoEPermuteExpertsUnpermute):
         """
         workspace1 = (M, K)
         workspace2 = (0,)
-        output_shape = (M, K * 2 if self.quant_dtype == "nvfp4" else K)
+        # For TP, the quantization is fused with fused_moe call.
+        output_shape = (M, K * 2 if self.quant_dtype == "nvfp4" and self.use_dp else K)
         # The workspace is determined by `aq`, since it comes after any
         # potential communication op and is involved in the expert computation.
         return (workspace1, workspace2, output_shape)
@@ -214,6 +217,7 @@ def flashinfer_cutlass_moe_fp4(
         FlashInferExperts(
             out_dtype=hidden_states.dtype,
             quant_config=quant_config,
+            use_dp=False,
         ),
     )
 

vllm/model_executor/layers/fused_moe/flashinfer_cutlass_prepare_finalize.py

@@ -170,6 +170,8 @@ class FlashInferAllGatherMoEPrepareAndFinalize(FlashInferCutlassMoEPrepareAndFin
         self._apply_router_weight_on_input(
             a1, topk_weights, topk_ids, apply_router_weight_on_input
         )
+        if not self.use_dp:
+            return a1, None, None, topk_ids, topk_weights
 
         a1q, a1q_scale = moe_kernel_quantize_input(
             a1,
@@ -179,7 +181,6 @@ class FlashInferAllGatherMoEPrepareAndFinalize(FlashInferCutlassMoEPrepareAndFin
             quant_config.block_shape,
             is_fp4_scale_swizzled=not self.use_dp,
         )
-        if self.use_dp:
         topk_weights, topk_ids, a1q, a1q_scale = get_dp_group().all_gatherv(
             [topk_weights, topk_ids, a1q, a1q_scale],
             dim=0,

vllm/model_executor/layers/quantization/modelopt.py

@@ -1769,29 +1769,6 @@ class ModelOptNvFp4FusedMoE(FusedMoEMethodBase):
                 expert_map=expert_map,
                 apply_router_weight_on_input=apply_router_weight_on_input,
             )
-        elif (
-            self.allow_flashinfer
-            and self.flashinfer_moe_backend == FlashinferMoeBackend.CUTLASS
-        ):
-            from vllm.model_executor.layers.fused_moe.flashinfer_cutlass_moe import (  # noqa: E501
-                flashinfer_cutlass_moe_fp4,
-            )
-
-            assert self.moe_quant_config is not None
-
-            return flashinfer_cutlass_moe_fp4(
-                hidden_states=x,
-                w1=layer.w13_weight,
-                w2=layer.w2_weight,
-                topk_weights=topk_weights,
-                topk_ids=topk_ids,
-                quant_config=self.moe_quant_config,
-                inplace=False,
-                activation=activation,
-                global_num_experts=global_num_experts,
-                expert_map=expert_map,
-                apply_router_weight_on_input=apply_router_weight_on_input,
-            )
         else:
             # If no modular kernel is provided, use cutlass_moe_fp4 for TP case
             # only (no EP).

vllm/model_executor/layers/quantization/utils/flashinfer_fp4_moe.py

@@ -79,6 +79,7 @@ def select_nvfp4_gemm_impl(
             ep_size=moe.moe_parallel_config.ep_size,
             tp_rank=moe.moe_parallel_config.tp_rank,
             tp_size=moe.moe_parallel_config.tp_size,
+            use_dp=moe.moe_parallel_config.dp_size > 1,
         )
 
     # native cutlass experts currently don't support DP; TP case won't call this