[SW-226128]Disable mark scales as const & add dist.barrier only for multiple cards

README_GAUDI.md

@@ -409,6 +409,9 @@ measurements for a given model. The quantization configuration is used during in
 > If you are prototyping or testing your model with FP8, you can use the `VLLM_SKIP_WARMUP=true` environment variable to disable the warmup stage, which is time-consuming.
 However, disabling this feature in production environments is not recommended, as it can lead to a significant performance decrease.
 
+> [!TIP]
+> If you are benchmarking an FP8 model with `scale_format=const`, setting `VLLM_DISABLE_MARK_SCALES_AS_CONST=true` can help speed up the warmup stage.
+
 > [!TIP]
 > When using FP8 models, you may experience timeouts caused by the long compilation time of FP8 operations. To mitigate this, set the following environment variables:
 > - `VLLM_ENGINE_ITERATION_TIMEOUT_S` - to adjust the vLLM server timeout. You can set the value in seconds, e.g., 600 equals 10 minutes.

vllm/worker/hpu_model_runner.py

@@ -825,16 +825,22 @@ def load_model(self) -> None:
                 with HabanaMemoryProfiler() as m_inc:
                     from neural_compressor.torch.quantization import (
                         FP8Config, convert, prepare)
+
+                    disable_mark_scales_as_const = os.getenv(
+                        "VLLM_DISABLE_MARK_SCALES_AS_CONST",
+                        "false") in ("1", "true")
                     config = FP8Config.from_json_file(
                         os.getenv("QUANT_CONFIG", ""))
                     self._inc_preprocess_(self.model, config)
                     if config.measure:
                         self.model = prepare(self.model, config)
                     elif config.quantize:
                         self.model = convert(self.model, config)
-                    htcore.hpu_initialize(self.model,
-                                          mark_only_scales_as_const=True)
-                    torch.distributed.barrier()
+                    if not disable_mark_scales_as_const:
+                        htcore.hpu_initialize(self.model,
+                                              mark_only_scales_as_const=True)
+                    if torch.distributed.is_initialized():
+                        torch.distributed.barrier()
                 self.inc_initialized_successfully = True
                 logger.info("Preparing model with INC took %s",
                             m_inc.get_summary_string())