A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 5, 06:00 UTC
[compile] Enable AOT compile with batch invariance mode. (#39201)
Signed-off-by: zhxchen17 <zhxchen17@fb.com> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Zhengxu Chen committed
Apr 12, 2026 at 22:58 UTC
cae984060f6ff07745fe7a60b8cd0b81c98434b2
1 file changed
+1
-4
vllm/envs.py
+1
-4
@@ -300,10 +300,7 @@ def use_aot_compile() -> bool:
300
else "0"
301
)
302
303
- return (
304
- not bool(int(os.getenv("VLLM_BATCH_INVARIANT", "0")))
305
- and os.environ.get("VLLM_USE_AOT_COMPILE", default_value) == "1"
306
- )
303
+ return os.environ.get("VLLM_USE_AOT_COMPILE", default_value) == "1"
304
305
306
def use_mega_aot_artifact():