Skip to content

Revert "[Blackwell] Make mxint4 flashinfer_trtllm moe gemm set by default on blackwell" #19695

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
vincentzed wants to merge 1 commit into from
Closed

Revert "[Blackwell] Make mxint4 flashinfer_trtllm moe gemm set by default on blackwell" #19695

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 4 additions & 26 deletions python/sglang/srt/server_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1284,20 +1284,6 @@ def _handle_model_specific_adjustments(self):
# Set moe backend for DeepSeek
if is_sm100_supported():
quant_method = get_quantization_config(hf_config)
quant_cfg = getattr(hf_config, "quantization_config", None) or {}
config_groups = quant_cfg.get("config_groups", {})
group0 = config_groups.get("group_0", {})
weights_cfg = group0.get("weights", {})
# this also apply to kimi k2.5
# since it follow the compressed tensor int4 recipe
# but not kimi k2 instruct or 0905 instruct.
is_kimi_k2_k25_thinking_int4 = (
quant_method == "compressed-tensors"
and weights_cfg.get("num_bits") == 4
and weights_cfg.get("group_size") == 32
and weights_cfg.get("strategy") == "group"
and weights_cfg.get("type") == "int"
)
if self.quantization is None:
# Default DeepSeek V3/R1 native FP8 when not explicitly set,
# Because we need this condition for an assertion in
Expand All @@ -1312,20 +1298,12 @@ def _handle_model_specific_adjustments(self):
if (
self.moe_a2a_backend == "none"
and self.moe_runner_backend == "auto"
and (
self.quantization in ["fp8", "modelopt_fp8", "modelopt_fp4"]
or is_kimi_k2_k25_thinking_int4
)
and self.quantization in ["fp8", "modelopt_fp8", "modelopt_fp4"]
):
self.moe_runner_backend = "flashinfer_trtllm"
if is_kimi_k2_k25_thinking_int4:
logger.info(
"Use flashinfer_trtllm as MoE runner backend on Blackwell for Kimi K2 / K2.5 thinking int4"
)
else:
logger.info(
"Use flashinfer_trtllm as MoE runner backend on sm100 for DeepseekV3ForCausalLM"
)
logger.info(
"Use flashinfer_trtllm as MoE runner backend on sm100 for DeepseekV3ForCausalLM"
)
elif is_hip():
if not self.enable_dp_attention and self.nnodes == 1:
# TODO (Hubert): Put this back later
Expand Down
Loading