fix missing enable_pdl argument in trtllm-gen fp4 moe (#1480)

<!-- .github/pull_request_template.md -->

## 📌 Description

Fix the missing `enable_pdl` argument introduced in #1446 .

## 🔍 Related Issues

<!-- Link any related issues here -->

## 🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull
request, please make sure the following items are complete.

### ✅ Pre-commit Checks

- [x] I have installed `pre-commit` by running `pip install pre-commit`
(or used your preferred method).
- [x] I have installed the hooks with `pre-commit install`.
- [x] I have run the hooks manually with `pre-commit run --all-files`
and fixed any reported issues.

> If you are unsure about how to set up `pre-commit`, see [the
pre-commit documentation](https://pre-commit.com/).

## 🧪 Tests

- [ ] Tests have been added or updated as needed.
- [ ] All tests are passing (`unittest`, etc.).

## Reviewer Notes

<!-- Optional: anything you'd like reviewers to focus on, concerns, etc.
-->

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Avery Yingyi Huang <[email protected]>

Author: 3 people

flashinfer/fused_moe/core.py

@@ -593,6 +593,7 @@ def _fake_cutlass_fused_moe_sm100(
         use_mxfp8_act_scaling: bool = False,
         min_latency_mode: bool = False,
         tune_max_num_tokens: int = 8192,
+        enable_pdl: Optional[bool] = None,
     ):
         seq_len = input.shape[0]
         hidden_size = fc2_expert_weights.shape[1]
@@ -947,6 +948,7 @@ def _fake_trtllm_fp8_per_tensor_scale_moe(
         use_routing_scales_on_input: bool,
         tile_tokens_dim: int = 8,
         routing_method_type: int = 0,
+        enable_pdl: Optional[bool] = None,
     ):
         seq_len = hidden_states.shape[0]
         hidden_size = hidden_states.shape[1]
@@ -1031,6 +1033,7 @@ def _fake_trtllm_fp8_block_scale_moe(
         routing_method_type: int = 0,
         use_shuffled_weight: bool = False,
         weight_layout: int = 0,
+        enable_pdl: Optional[bool] = None,
     ):
         seq_len = hidden_states.shape[0]
         hidden_size = hidden_states.shape[1]
@@ -1174,6 +1177,8 @@ def _fake_trtllm_fp4_block_scale_moe(
         tile_tokens_dim: int,
         routing_method_type: int,
         do_finalize: bool,
+        enable_pdl: Optional[bool] = None,
+        output: Optional[torch.Tensor] = None,
     ):
         seq_len = hidden_states.shape[0]
         hidden_size = hidden_states.shape[1]
@@ -1231,6 +1236,7 @@ def trtllm_fp8_per_tensor_scale_moe(
         use_routing_scales_on_input: Whether to use routing scales on input
         tile_tokens_dim: Tile dimension for tokens (default: 8)
         routing_method_type: Type of routing method to use (default: 0)
+        enable_pdl: Whether to enable Programmatic Dependent Launch (PDL). Auto-enabled for >= sm90.
 
     Returns:
         torch.Tensor: Output tensor of shape [seq_len, hidden_size]
@@ -1303,7 +1309,7 @@ def trtllm_fp8_block_scale_moe(
         routed_scaling_factor: Scaling factor for routing
         tile_tokens_dim: Tile dimension for tokens (default: 8)
         routing_method_type: Type of routing method to use (default: 0)
-
+        enable_pdl: Whether to enable Programmatic Dependent Launch (PDL). Auto-enabled for >= sm90.
     Returns:
         torch.Tensor: Output tensor of shape [seq_len, hidden_size]
     """
@@ -1360,6 +1366,7 @@ def trtllm_fp4_block_scale_moe(
     tile_tokens_dim: int = 8,
     routing_method_type: int = 0,
     do_finalize: bool = True,
+    enable_pdl: Optional[bool] = None,
     output: Optional[torch.Tensor] = None,
 ) -> List[torch.Tensor]:
     """FP4 block scale MoE operation.
@@ -1405,7 +1412,7 @@ def trtllm_fp4_block_scale_moe(
         do_finalize (bool): Whether to finalize the output (default: False)
         output (Optional[torch.Tensor]): shape [seq_len, hidden_size]
             Optional inplace output tensor.
-
+        enable_pdl: Whether to enable Programmatic Dependent Launch (PDL). Auto-enabled for >= sm90.
     Returns:
         List[torch.Tensor]: List of output tensors. If do_finalize=True, returns the final MoE output.
             Otherwise, returns intermediate results (gemm2_output, expert_weights, expanded_idx_to_permuted_idx) that need further processing.
@@ -1440,6 +1447,7 @@ def trtllm_fp4_block_scale_moe(
         tile_tokens_dim,
         routing_method_type,
         do_finalize,
+        enable_pdl,
         output,
     )
 

tests/test_trtllm_gen_fused_moe.py

@@ -709,6 +709,7 @@ def call_moe(
         routed_scaling = kwargs["routed_scaling"]
         routing_method_type = kwargs["routing_method_type"]
         tile_tokens_dim = kwargs["tile_tokens_dim"]
+        enable_pdl = kwargs.get("enable_pdl")
 
         # Generate block scales and quantize hidden states at runtime
         hidden_states_fp8 = hidden_states_orig.to(torch.float8_e4m3fn)
@@ -738,6 +739,7 @@ def call_moe(
             routing_method_type,
             use_shuffled_weight=static_data["use_shuffled_weight"],
             weight_layout=static_data["weight_layout"],
+            enable_pdl=enable_pdl,
         )
 
         return output.to(torch.float)
@@ -2040,6 +2042,7 @@ def test_moe_quantization_classes(
         routing_method_type=routing_method_type,
         tile_tokens_dim=tile_tokens_dim,
         weight_processing=weight_processing,
+        enable_pdl=True,
     )
 
     # Compare outputs using moe_impl-specific tolerances