Raise error for data-parallel with benchmark_throughput (vllm-project#16737)

Signed-off-by: Kartik Ramesh <kartikx2000@gmail.com>
Co-authored-by: Simon Mo <simon.mo@hey.com>
Signed-off-by: dtransposed <damian@damian-ml-machine.europe-west3-b.c.jetbrains-grazie.internal>

Author: 2 people

benchmarks/benchmark_throughput.py

@@ -523,6 +523,13 @@ def validate_args(args):
         raise ValueError(
             "Tokenizer must be the same as the model for MII backend.")
 
+    # --data-parallel is not supported currently.
+    # https://github.com/vllm-project/vllm/issues/16222
+    if args.data_parallel_size > 1:
+        raise ValueError(
+            "Data parallel is not supported in offline benchmark, \
+            please use benchmark serving instead")
+
 
 if __name__ == "__main__":
     parser = FlexibleArgumentParser(description="Benchmark the throughput.")