[2/2] Add python wrapper for CUTLASS FP8 Blockscale MoE Kernel. (#5694)

python/sglang/srt/layers/quantization/fp8_utils.py

@@ -80,6 +80,12 @@ def cutlass_fp8_supported():
     return False
 
 
+def is_sm100_supported(device=None) -> bool:
+    return (torch.cuda.get_device_capability(device)[0] == 10) and (
+        torch.version.cuda >= "12.8"
+    )
+
+
 def normalize_e4m3fn_to_e4m3fnuz(
     weight: torch.Tensor,
     weight_scale: torch.Tensor,