[Refactor] Cuda Graph Runner/Backend Refactor (#23906)

Co-authored-by: BBuf <1182563586@qq.com>
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>
Co-authored-by: Lianmin Zheng <lianminzheng@gmail.com>

python/sglang/test/runners.py

@@ -578,7 +578,7 @@ class SRTRunner:
         disable_overlap_schedule: bool = False,
         disable_custom_all_reduce: bool = False,
         torchao_config: Optional[str] = None,
-        cuda_graph_max_bs: int = 4,
+        cuda_graph_max_bs_decode: int = 4,
         sleep_on_idle=False,
         max_lora_rank: Optional[int] = None,
         lora_target_modules: Optional[List[str]] = None,
@@ -635,7 +635,7 @@ class SRTRunner:
             dp_size=dp_size,
             tokenizer_path=tokenizer_path,
             disable_overlap_schedule=disable_overlap_schedule,
-            cuda_graph_max_bs=cuda_graph_max_bs,
+            cuda_graph_max_bs_decode=cuda_graph_max_bs_decode,
             disable_custom_all_reduce=disable_custom_all_reduce,
             sleep_on_idle=sleep_on_idle,
             max_lora_rank=max_lora_rank,