[PD] Support prefill overlap + Ensure no race condition (#5609)

This commit is contained in:
Byron Hsu
2025-04-21 12:12:56 -07:00
committed by GitHub
parent e65b9f21e3
commit bf98d2e377
5 changed files with 107 additions and 18 deletions

View File

@@ -388,8 +388,6 @@ class ServerArgs:
if self.disaggregation_mode == "prefill":
self.disable_cuda_graph = True
logger.warning("Cuda graph is disabled for prefill server")
self.disable_overlap_schedule = True
logger.warning("Overlap scheduler is disabled for prefill server")
elif self.disaggregation_mode == "decode":
self.disable_radix_cache = True
logger.warning("KV cache is forced as chunk cache for decode server")