move all get_stream in sgl_kernel to c++ to reduce the launch overhead (#12521)

This commit is contained in:
Lianmin Zheng
2025-11-02 13:15:05 -08:00
committed by GitHub
parent c9db79117f
commit 20315697f4
20 changed files with 61 additions and 94 deletions

View File

@@ -229,7 +229,6 @@ class Envs:
SGLANG_SKIP_SGL_KERNEL_VERSION_CHECK = EnvBool(False)
# vLLM dependencies (TODO: they have been deprecated, we can remove them safely)
USE_VLLM_CUSTOM_ALLREDUCE = EnvBool(False)
USE_VLLM_CUTLASS_W8A8_FP8_KERNEL = EnvBool(False)
USE_TRITON_W8A8_FP8_KERNEL = EnvBool(False)