[Core/DBO][1/N] Add Dual-Batch Overlap mechanism to VLLM (#23693)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Sage Moore <sage@neuralmagic.com>
Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Co-authored-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>

examples/offline_inference/data_parallel.py

@@ -87,6 +87,11 @@ def parse_args():
         default=0.8,
         help=("Fraction of GPU memory vLLM is allowed to allocate (0.0, 1.0]."),
     )
+    parser.add_argument(
+        "--enable-dbo",
+        action="store_true",
+        help=("Enable microbatched execution"),
+    )
     parser.add_argument(
         "--compilation-config",
         type=int,
@@ -113,6 +118,7 @@ def main(
     max_model_len,
     compilation_config,
     gpu_memory_utilization,
+    enable_dbo,
     quantization,
 ):
     os.environ["VLLM_DP_RANK"] = str(global_dp_rank)
@@ -167,6 +173,7 @@ def main(
         max_num_seqs=max_num_seqs,
         max_model_len=max_model_len,
         gpu_memory_utilization=gpu_memory_utilization,
+        enable_dbo=enable_dbo,
         quantization=quantization,
         compilation_config=compilation_config,
     )
@@ -227,6 +234,7 @@ if __name__ == "__main__":
                 args.max_model_len,
                 args.compilation_config,
                 args.gpu_memory_utilization,
+                args.enable_dbo,
                 args.quantization,
             ),
         )