EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Refactor] Migrate profiler config from env vars to explicit ProfilerConfig (#5928)

Browse Source 
### What this PR does / why we need it?

Migrate the torch profiler configuration from deprecated environment
variables (`VLLM_TORCH_PROFILER_DIR`, `VLLM_TORCH_PROFILER_WITH_STACK`,
`VLLM_TORCH_PROFILER_WITH_PROFILE_MEMORY`) to the explicit
`ProfilerConfig` object, aligning with vLLM's configuration best
practices.
The profiler environment variable approach is deprecated in vLLM and
will be removed in v0.14.0 or v1.0.0.

### Does this PR introduce _any_ user-facing change?
yes, for deverlopers who want to fetch profiler, he should use `--profiler-config` instead of `VLLM_TORCH_PROFILER_DIR`
### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
11b6af5280

Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
This commit is contained in:
meihanc
2026-01-19 09:27:55 +08:00
committed by GitHub
parent bc1f6713e7
commit 9cad1a8349
3 changed files with 68 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
docs/source/tutorials/DeepSeek-V3.2.md
View File

@@ -278,9 +278,6 @@ Before you start, please
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256
export VLLM_TORCH_PROFILER_DIR="./vllm_profile"
export VLLM_TORCH_PROFILER_WITH_STACK=0
export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
@@ -302,6 +299,10 @@ Before you start, please
--tensor-parallel-size $7 \
--enable-expert-parallel \
--speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
--profiler-config \
'{"profiler": "torch",
"torch_profiler_dir": "./vllm_profile",
"torch_profiler_with_stack": false}' \
--seed 1024 \
--served-model-name dsv3 \
--max-model-len 68000 \
@@ -351,9 +352,6 @@ Before you start, please
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256
export VLLM_TORCH_PROFILER_DIR="./vllm_profile"
export VLLM_TORCH_PROFILER_WITH_STACK=0
export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
@@ -376,6 +374,10 @@ Before you start, please
--tensor-parallel-size $7 \
--enable-expert-parallel \
--speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
--profiler-config \
'{"profiler": "torch",
"torch_profiler_dir": "./vllm_profile",
"torch_profiler_with_stack": false}' \
--seed 1024 \
--served-model-name dsv3 \
--max-model-len 68000 \
@@ -426,8 +428,6 @@ Before you start, please
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256
export VLLM_TORCH_PROFILER_DIR="./vllm_profile"
export VLLM_TORCH_PROFILER_WITH_STACK=0
export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
@@ -452,6 +452,10 @@ Before you start, please
--tensor-parallel-size $7 \
--enable-expert-parallel \
--speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
--profiler-config \
'{"profiler": "torch",
"torch_profiler_dir": "./vllm_profile",
"torch_profiler_with_stack": false}' \
--seed 1024 \
--served-model-name dsv3 \
--max-model-len 68000 \
@@ -504,9 +508,6 @@ Before you start, please
export VLLM_USE_V1=1
export HCCL_BUFFSIZE=256
export VLLM_TORCH_PROFILER_DIR="./vllm_profile"
export VLLM_TORCH_PROFILER_WITH_STACK=0
export ASCEND_AGGREGATE_ENABLE=1
export ASCEND_TRANSPORT_PRINT=1
export ACL_OP_INIT_MODE=1
@@ -530,6 +531,10 @@ Before you start, please
--tensor-parallel-size $7 \
--enable-expert-parallel \
--speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}' \
--profiler-config \
'{"profiler": "torch",
"torch_profiler_dir": "./vllm_profile",
"torch_profiler_with_stack": false}' \
--seed 1024 \
--served-model-name dsv3 \
--max-model-len 68000 \