[Bugfix][ROCm] Fix worker startup OOM on ROCm by skipping unreliable cudagraph...

[Bugfix][ROCm] Fix worker startup OOM on ROCm by skipping unreliable cudagraph memory profiling (#36720)
Signed-off-by: JartX <sagformas@epdcenter.es>
(cherry picked from commit e8f9dbc3)

vllm/v1/worker/gpu_worker.py

@@ -392,8 +392,10 @@ class Worker(WorkerBase):
             )
 
             # Profile CUDA graph memory if graphs will be captured.
+            # Skip on ROCm/HIP as graph pool handles and mem_get_info behave
+            # differently and can produce incorrect/negative estimates.
             cudagraph_memory_estimate = 0
-            if not self.model_config.enforce_eager:
+            if not self.model_config.enforce_eager and not current_platform.is_rocm():
                 cudagraph_memory_estimate = self.model_runner.profile_cudagraph_memory()
 
         # Use the pre-cudagraph torch peak to avoid double-counting.
@@ -406,6 +408,8 @@ class Worker(WorkerBase):
             + profile_result.weights_memory
         )
 
+        # On ROCm, cudagraph_memory_estimate is always 0 so this is a no-op.
+        # On CUDA, respect the opt-in flag as originally designed.
         cudagraph_memory_estimate_applied = (
             cudagraph_memory_estimate
             if envs.VLLM_MEMORY_PROFILER_ESTIMATE_CUDAGRAPHS
@@ -517,7 +521,6 @@ class Worker(WorkerBase):
 
     def update_max_model_len(self, max_model_len: int) -> None:
         """Update max_model_len after auto-fit to GPU memory.
-
         This is called when max_model_len=-1 is used and the engine
         automatically determines the maximum context length that fits
         in GPU memory. Workers need to update their cached max_model_len