[Bugfix] Respect cgroup memory limits on all platforms (#49966)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>

vllm/model_executor/model_loader/weight_utils.py

@@ -694,12 +694,10 @@ def _get_checkpoints_size_bytes(files: list[str]) -> int:
 
 
 def _get_available_ram_bytes() -> int:
-    """Return available RAM, honoring cgroup limits on ROCm."""
+    """Return available RAM, honoring cgroup limits."""
     import psutil
 
     host_available = psutil.virtual_memory().available
-    if not current_platform.is_rocm():
-        return host_available
 
     from vllm.utils.cpu_resource_utils import get_cgroup_memory_limit