EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Improve profiler and integrate profiler in bench_one_batch_server (#6787)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-05-31 15:53:55 -07:00
committed by GitHub
parent b520d02888
commit 2d72fc47cf
25 changed files with 481 additions and 223 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
python/sglang/srt/model_executor/cuda_graph_runner.py
View File

@@ -39,10 +39,7 @@ from sglang.srt.model_executor.forward_batch_info import (
PPProxyTensors,
)
from sglang.srt.patch_torch import monkey_patch_torch_compile
from sglang.srt.two_batch_overlap import (
TboCudaGraphRunnerPlugin,
TboForwardBatchPreparer,
)
from sglang.srt.two_batch_overlap import TboCudaGraphRunnerPlugin
from sglang.srt.utils import (
get_available_gpu_memory,
get_device_memory_capacity,

6
python/sglang/srt/model_executor/model_runner.py
View File

@@ -77,11 +77,7 @@ from sglang.srt.model_executor.cuda_graph_runner import CudaGraphRunner
from sglang.srt.model_executor.expert_location_updater import ExpertLocationUpdater
from sglang.srt.model_executor.forward_batch_info import ForwardBatch, PPProxyTensors
from sglang.srt.model_loader import get_model
from sglang.srt.model_loader.loader import (
DefaultModelLoader,
device_loading_context,
get_model_loader,
)
from sglang.srt.model_loader.loader import DefaultModelLoader, get_model_loader
from sglang.srt.model_loader.utils import set_default_torch_dtype
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.patch_torch import monkey_patch_torch_reductions