[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>

test/registered/model_loading/test_prefetch_checkpoints_multi_gpu.py

@@ -26,7 +26,7 @@ class TestPrefetchCheckpointsMultiGPU(CustomTestCase):
             enable_dp_attention=True,
             disable_radix_cache=True,
             weight_loader_prefetch_checkpoints=True,
-            cuda_graph_max_bs=1,
+            cuda_graph_max_bs_decode=1,
             max_total_tokens=256,
         )