[misc] Use --cuda-graph-max-bs-decode in tests, examples, and docs (#29591)

test/manual/dsv4/test_h200_fp8_pro.py

@@ -44,7 +44,7 @@ class TestH200Fp8ProLowLatency(DSV4ProAime25TestBase):
             *multinode_args(2),
             "--moe-a2a-backend",
             "deepep",
-            "--cuda-graph-max-bs",
+            "--cuda-graph-max-bs-decode",
             "8",
             "--max-running-requests",
             "32",
@@ -88,7 +88,7 @@ class TestH200Fp8ProBalanced(DSV4ProAime25TestBase):
             "2",
             "--mem-fraction-static",
             "0.88",
-            "--cuda-graph-max-bs",
+            "--cuda-graph-max-bs-decode",
             "8",
             "--max-running-requests",
             "32",
@@ -114,7 +114,7 @@ class TestH200Fp8ProMaxThroughput(DSV4ProAime25TestBase):
             "deepep",
             "--mem-fraction-static",
             "0.88",
-            "--cuda-graph-max-bs",
+            "--cuda-graph-max-bs-decode",
             "128",
             "--max-running-requests",
             "256",