Fix MoE backend selection for LoRA (unquantized MoE)

tests/kernels/moe/test_unquantized_backend_selection.py

@@ -11,6 +11,11 @@
 )
 from vllm.platforms import current_platform
 
+skipif_not_cuda_rocm = pytest.mark.skipif(
+    not (current_platform.is_cuda() or current_platform.is_rocm()),
+    reason="Only supported on CUDA/ROCm platforms.",
+)
+
 
 @pytest.mark.parametrize(
     "platform_method,expected_backend",
@@ -190,3 +195,83 @@ def test_select_cuda_flashinfer_cutlass_backend(
 
         assert selected_backend == UnquantizedMoeBackend.FLASHINFER_CUTLASS
         assert experts_cls is not None
+
+
+@skipif_not_cuda_rocm
+def test_select_lora_backend_prefers_triton():
+    """LoRA-enabled unquantized MoE should select Triton backend."""
+    moe_config = make_dummy_moe_config()
+    moe_config.is_lora_enabled = True
+    selected_backend, experts_cls = select_unquantized_moe_backend(
+        moe_config=moe_config
+    )
+
+    assert selected_backend == UnquantizedMoeBackend.TRITON
+    assert experts_cls is not None
+
+
+@skipif_not_cuda_rocm
+def test_select_lora_explicit_non_triton_backend():
+    """LoRA should override explicit non-Triton backend to Triton."""
+    moe_config = make_dummy_moe_config()
+    moe_config.is_lora_enabled = True
+
+    # Use string from mapping in function map_unquantized_backend()
+    moe_config.moe_backend = "flashinfer_cutlass"
+
+    selected_backend, experts_cls = select_unquantized_moe_backend(
+        moe_config=moe_config
+    )
+
+    assert selected_backend == UnquantizedMoeBackend.TRITON
+    assert experts_cls is not None
+
+
+@skipif_not_cuda_rocm
+@pytest.mark.parametrize("is_lora_enabled", [False, True])
+def test_select_explicit_triton_backend(is_lora_enabled):
+    """Explicit triton backend selection should return Triton."""
+    moe_config = make_dummy_moe_config()
+    moe_config.is_lora_enabled = is_lora_enabled
+    moe_config.moe_backend = "triton"
+
+    selected_backend, experts_cls = select_unquantized_moe_backend(
+        moe_config=moe_config
+    )
+
+    assert selected_backend == UnquantizedMoeBackend.TRITON
+    assert experts_cls is not None
+
+
+@skipif_not_cuda_rocm
+def test_select_explicit_triton_ignores_flashinfer_env(monkeypatch):
+    """Explicit triton backend should override FlashInfer env selection."""
+    monkeypatch.setenv("VLLM_USE_FLASHINFER_MOE_FP16", "1")
+    monkeypatch.setenv("VLLM_FLASHINFER_MOE_BACKEND", "throughput")
+
+    moe_config = make_dummy_moe_config()
+    moe_config.is_lora_enabled = False
+    moe_config.moe_backend = "triton"
+
+    selected_backend, experts_cls = select_unquantized_moe_backend(
+        moe_config=moe_config
+    )
+
+    assert selected_backend == UnquantizedMoeBackend.TRITON
+    assert experts_cls is not None
+
+
+@skipif_not_cuda_rocm
+def test_select_lora_ignores_flashinfer_env(monkeypatch):
+    """LoRA path should still choose Triton even if FlashInfer env is on."""
+    monkeypatch.setenv("VLLM_USE_FLASHINFER_MOE_FP16", "1")
+    monkeypatch.setenv("VLLM_FLASHINFER_MOE_BACKEND", "throughput")
+
+    moe_config = make_dummy_moe_config()
+    moe_config.is_lora_enabled = True
+    selected_backend, experts_cls = select_unquantized_moe_backend(
+        moe_config=moe_config
+    )
+
+    assert selected_backend == UnquantizedMoeBackend.TRITON
+    assert experts_cls is not None

vllm/model_executor/layers/fused_moe/oracle/unquantized.py

@@ -163,6 +163,11 @@ def select_unquantized_moe_backend(
     if current_platform.is_out_of_tree():
         return UnquantizedMoeBackend.OOT, None
 
+    if moe_config.is_lora_enabled:
+        return UnquantizedMoeBackend.TRITON, backend_to_kernel_cls(
+            UnquantizedMoeBackend.TRITON
+        )
+
     # NOTE: the kernels are selected in the following order.
     AVAILABLE_BACKENDS = _get_priority_backends(moe_config)