[Elastic EP] Preserve AOT cache reuse during scaling (#53378)
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
Itay Alroy committed
Aug 27, 2026 at 17:55 UTC
6d1112260739e26c18a9176376fe1dd4ad03f90d
2 files changed
+3
vllm/config/compilation.py
+2
@@ -798,6 +798,8 @@ class CompilationConfig:
798
"traced_files",
799
"compilation_time",
800
"encoder_compilation_time",
801
+ "enabled_custom_ops",
802
+ "disabled_custom_ops",
803
"static_forward_context",
804
"pass_config", # handled separately below
805
"dynamic_shapes_config", # handled separately below
vllm/envs.py
+1
@@ -2270,6 +2270,7 @@ def compile_factors() -> dict[str, object]:
2270
"VLLM_RANDOMIZE_DP_DUMMY_INPUTS",
2271
"VLLM_MODEL_REDIRECT_PATH",
2272
"VLLM_HOST_IP",
2273
+ "VLLM_ELASTIC_EP_SCALE_UP_LAUNCH",
2274
"VLLM_FORCE_AOT_LOAD",
2275
"S3_ACCESS_KEY_ID",
2276
"S3_SECRET_ACCESS_KEY",