Support triton_kernels for GPT-OSS on SM120

python/sglang/srt/layers/quantization/mxfp4.py

@@ -37,12 +37,12 @@
 from sglang.srt.layers.quantization.utils import is_layer_skipped
 from sglang.srt.server_args import get_global_server_args
 from sglang.srt.utils import (
-    is_cuda,
     is_flashinfer_available,
     is_gfx95_supported,
     is_hip,
     is_sm90_supported,
     is_sm100_supported,
+    is_sm120_supported,
     is_triton_kernels_available,
     mxfp_supported,
     next_power_of_2,
@@ -52,8 +52,6 @@
 from sglang.srt.utils.common import get_bool_env_var
 from sglang.srt.utils.custom_op import register_custom_op
 
-_is_sm100_supported = is_cuda() and is_sm100_supported()
-_is_sm90_supported = is_cuda() and is_sm90_supported()
 has_triton_kernels = is_triton_kernels_available()
 
 
@@ -140,23 +138,40 @@ def _swizzle_mxfp4(quant_tensor, scale, num_warps):
     from triton_kernels.tensor import FP4, convert_layout, wrap_torch_tensor
     from triton_kernels.tensor_details import layout
 
-    value_layout, value_layout_opts = layout.make_default_matmul_mxfp4_w_layout(
-        mx_axis=1
-    )
-    scale_layout, scale_layout_opts = layout.make_default_matmul_mxfp4_w_scale_layout(
-        mx_axis=1, num_warps=num_warps
-    )
-    if _is_sm100_supported:
-        constraints = {
-            "is_persistent": True,
-            "epilogue_subtile": 1,
-        }
-        opt_flags.update_opt_flags_constraints(constraints)
-    elif _is_sm90_supported:
+    if is_sm120_supported():
+        # SM120 (Blackwell desktop) doesn't support persistent kernels / TMA block layout
+        # Use StridedLayout and disable persistent kernels to avoid assertion errors
+        from triton_kernels.tensor_details.layout import StridedLayout
+
+        value_layout = StridedLayout
+        value_layout_opts = {}
+        scale_layout = StridedLayout
+        scale_layout_opts = {}
         constraints = {
-            "split_k": 1,
+            "is_persistent": False,
+            "num_stages": 1,
         }
         opt_flags.update_opt_flags_constraints(constraints)
+    else:
+        value_layout, value_layout_opts = layout.make_default_matmul_mxfp4_w_layout(
+            mx_axis=1
+        )
+        scale_layout, scale_layout_opts = (
+            layout.make_default_matmul_mxfp4_w_scale_layout(
+                mx_axis=1, num_warps=num_warps
+            )
+        )
+        if is_sm100_supported():
+            constraints = {
+                "is_persistent": True,
+                "epilogue_subtile": 1,
+            }
+            opt_flags.update_opt_flags_constraints(constraints)
+        elif is_sm90_supported():
+            constraints = {
+                "split_k": 1,
+            }
+            opt_flags.update_opt_flags_constraints(constraints)
     # transpose the tensor so that the quantization axis is on dim1
     quant_tensor = quant_tensor.transpose(-2, -1)
     scale = scale.transpose(-2, -1)
@@ -324,7 +339,7 @@ def create_weights(
         # pad the intermediate size to be a multiple of 2 * mxfp4_block
         # for to hold non-uniform sharded tensor as well as swizzling
         intermediate_size_per_partition_after_pad = intermediate_size_per_partition
-        if _is_sm100_supported:
+        if is_sm100_supported():
             if self.use_flashinfer:
                 intermediate_size_per_partition_after_pad = round_up(
                     intermediate_size_per_partition, 256

python/sglang/srt/server_args.py

@@ -1486,10 +1486,16 @@ def _handle_model_specific_adjustments(self):
                 self.dtype = "bfloat16"
 
             if self.moe_runner_backend == "auto":
-                if is_blackwell_supported() and is_mxfp4_quant_format:
+                if is_sm100_supported() and is_mxfp4_quant_format:
                     self.moe_runner_backend = "flashinfer_mxfp4"
                     logger.warning(
-                        "Detected Blackwell and MXFP4 quantization format for GPT-OSS model, enabling FlashInfer MXFP4 MOE kernel."
+                        "Detected SM100 and MXFP4 quantization format for GPT-OSS model, enabling FlashInfer MXFP4 MOE kernel."
+                    )
+                elif is_sm120_supported() and is_mxfp4_quant_format:
+                    # trtllm-gen only supports SM100
+                    self.moe_runner_backend = "triton_kernel"
+                    logger.warning(
+                        "Detected SM120 and MXFP4 quantization format for GPT-OSS model, enabling triton_kernel MOE kernel."
                     )
                 elif (
                     is_hip() and get_bool_env_var("SGLANG_USE_AITER")