refactor ux

Signed-off-by: Robert Shaw <robshaw@redhat.com>

vllm/engine/arg_utils.py

@@ -295,6 +295,7 @@ class EngineArgs:
     tensor_parallel_size: int = ParallelConfig.tensor_parallel_size
     data_parallel_size: int = ParallelConfig.data_parallel_size
     data_parallel_rank: Optional[int] = None
+    data_parallel_start_rank: Optional[int] = None
     data_parallel_size_local: Optional[int] = None
     data_parallel_address: Optional[str] = None
     data_parallel_rpc_port: Optional[int] = None
@@ -606,6 +607,11 @@ class EngineArgs:
             type=int,
             help='Data parallel rank of this instance. '
             'When set, enables external load balancer mode.')
+        parallel_group.add_argument('--data-parallel-start-rank',
+                                    '-dpr',
+                                    type=int,
+                                    help='Starting data parallel rank '
+                                    'for secondary nodes.')
         parallel_group.add_argument('--data-parallel-size-local',
                                     '-dpl',
                                     type=int,
@@ -1091,19 +1097,36 @@ class EngineArgs:
             # but we should not do this here.
             placement_group = ray.util.get_current_placement_group()
 
-        # data_parallel_external_lb = self.data_parallel_rank is not None
-        # if data_parallel_external_lb:
-        #     assert self.data_parallel_size_local in (1, None), (
-        #         "data_parallel_size_local must be 1 when data_parallel_rank "
-        #         "is set")
-        #     data_parallel_size_local = 1
-        # elif self.data_parallel_size_local is not None:
-        data_parallel_external_lb = False
-        if self.data_parallel_size_local is not None:
+        # Organize --data-parallel-start-rank and --data-parallel-rank.
+        if self.data_parallel_start_rank is not None:
+            if self.data_parallel_rank is not None:
+                raise ValueError(
+                    "Found --data-parallel-rank and --data-parallel-start-rank."
+                    "Only one should be set (use --data-parallel-start-rank).")
+            else:
+                self.data_parallel_rank = self.data_parallel_start_rank
+
+        # Validate External LB.
+        data_parallel_external_lb = True
+        if data_parallel_external_lb:
+            if self.data_parallel_size_local is None:
+                raise ValueError(
+                    "With external LB, --data-parallel-size-local must be set."
+                )
+            if self.data_parallel_size_local >= self.data_parallel_size:
+                raise ValueError(
+                    "With external LB, --data-parallel-size-local must be less "
+                    "than --data-parallel-size.")
+            if (self.data_parallel_rank is not None
+                    and self.data_parallel_size_local > 1):
+                raise ValueError(
+                    "With --data-parallel-size-local > 1, use --data-parall"
+                    "--data-parallel-rank")
             data_parallel_size_local = self.data_parallel_size_local
-        else:
-            # Local DP size defaults to global DP size if not set.
-            data_parallel_size_local = self.data_parallel_size
+
+        # Local DP size defaults to global DP size if not set.
+        data_parallel_size_local = (self.data_parallel_size_local
+                                    or self.data_parallel_size)
 
         # DP address, used in multi-node case for torch distributed group
         # and ZMQ sockets.

vllm/entrypoints/cli/serve.py

@@ -45,11 +45,6 @@ class ServeSubcommand(CLISubcommand):
         if args.headless or args.api_server_count < 1:
             run_headless(args)
         else:
-            if args.data_parallel_start_rank:
-                raise ValueError(
-                    "data_parallel_start_rank is only applicable "
-                    "in headless mode. "
-                    "Add --headless flag to enable headless mode.")
             if args.api_server_count > 1:
                 run_multi_api_server(args)
             else:
@@ -122,7 +117,7 @@ def run_headless(args: argparse.Namespace):
     engine_manager = CoreEngineProcManager(
         target_fn=EngineCoreProc.run_engine_core,
         local_engine_count=local_engine_count,
-        start_index=args.data_parallel_start_rank,
+        start_index=vllm_config.parallel_config.data_parallel_rank,
         local_start_index=0,
         vllm_config=vllm_config,
         local_client=False,

vllm/entrypoints/openai/cli_args.py

@@ -253,13 +253,6 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
         default=False,
         help="Run in headless mode. See multi-node data parallel "
         "documentation for more details.")
-    parser.add_argument(
-        "--data-parallel-start-rank",
-        "-dpr",
-        type=int,
-        default=0,
-        help="Starting data parallel rank for secondary nodes. "
-        "Requires --headless.")
     parser.add_argument("--api-server-count",
                         "-asc",
                         type=int,