Unverified Commit 1c472f8f authored by tmm77's avatar tmm77 Committed by GitHub
Browse files

Add get_device_uuid for rocm (#37694)


Signed-off-by: default avatarTiffany Mintz <Tiffany.Mintz@amd.com>
parent c57d38d6
...@@ -28,6 +28,7 @@ try: ...@@ -28,6 +28,7 @@ try:
from amdsmi import ( from amdsmi import (
AmdSmiException, AmdSmiException,
amdsmi_get_gpu_asic_info, amdsmi_get_gpu_asic_info,
amdsmi_get_gpu_device_uuid,
amdsmi_get_processor_handles, amdsmi_get_processor_handles,
amdsmi_init, amdsmi_init,
amdsmi_shut_down, amdsmi_shut_down,
...@@ -608,6 +609,20 @@ class RocmPlatform(Platform): ...@@ -608,6 +609,20 @@ class RocmPlatform(Platform):
return _ROCM_DEVICE_ID_NAME_MAP[device_name] return _ROCM_DEVICE_ID_NAME_MAP[device_name]
return asic_info["market_name"] return asic_info["market_name"]
@classmethod
@with_amdsmi_context
def get_device_uuid(cls, device_id: int = 0) -> str:
try:
device = amdsmi_get_processor_handles()[device_id]
except AmdSmiException as error:
logger.error("amdsmi device query failed ", exc_info=error)
return ""
try:
device_uuid = amdsmi_get_gpu_device_uuid(device)
except AmdSmiException as error:
logger.error("amdsmi device uuid query failed ", exc_info=error)
return device_uuid
@classmethod @classmethod
def get_device_total_memory(cls, device_id: int = 0) -> int: def get_device_total_memory(cls, device_id: int = 0) -> int:
device_props = torch.cuda.get_device_properties(device_id) device_props = torch.cuda.get_device_properties(device_id)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment