minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

fix: update Blackwell log/error messages to include SM12x (#18751)

Browse Source 
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
blake-snc
2026-02-16 00:23:51 +08:00
committed by GitHub
co-authored by Claude Opus 4.6
parent 07a24f1a38
commit 57f7e06cb9
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/server_args.py
+3 -3
View File
@@ -1207,7 +1207,7 @@ class ServerArgs:
if model_arch == "GlmMoeDsaForCausalLM" and is_blackwell_supported(): if model_arch == "GlmMoeDsaForCausalLM" and is_blackwell_supported():
envs.SGLANG_NSA_FORCE_MLA.set(True) envs.SGLANG_NSA_FORCE_MLA.set(True)
logger.warning( logger.warning(
"Force NSA prefill to use MLA (i.e. disable MHA_ONE_SHOT) for GlmMoeDsaForCausalLM on SM100." "Force NSA prefill to use MLA (i.e. disable MHA_ONE_SHOT) for GlmMoeDsaForCausalLM on Blackwell."
) )
if self.is_attention_backend_not_set(): if self.is_attention_backend_not_set():
self.attention_backend = "nsa" self.attention_backend = "nsa"
@@ -1372,7 +1372,7 @@ class ServerArgs:
if is_blackwell_supported() and is_mxfp4_quant_format: if is_blackwell_supported() and is_mxfp4_quant_format:
self.moe_runner_backend = "flashinfer_mxfp4" self.moe_runner_backend = "flashinfer_mxfp4"
logger.warning( logger.warning(
"Detected SM100 and MXFP4 quantization format for GPT-OSS model, enabling FlashInfer MXFP4 MOE kernel." "Detected Blackwell and MXFP4 quantization format for GPT-OSS model, enabling FlashInfer MXFP4 MOE kernel."
) )
elif ( elif (
is_hip() and get_bool_env_var("SGLANG_USE_AITER") is_hip() and get_bool_env_var("SGLANG_USE_AITER")
@@ -1874,7 +1874,7 @@ class ServerArgs:
): ):
if not is_blackwell_supported(): if not is_blackwell_supported():
raise ValueError( raise ValueError(
"TRTLLM MLA backend is only supported on Blackwell GPUs (SM100). Please use a different backend." "TRTLLM MLA backend is only supported on Blackwell GPUs (SM100/SM12x). Please use a different backend."
) )
if self.page_size not in [32, 64]: if self.page_size not in [32, 64]: