MCPcopy Create free account
hub / github.com/PaddlePaddle/Paddle / is_compiled_with_rocm

Function is_compiled_with_rocm

python/paddle/base/framework.py:1014–1027  ·  view source on GitHub ↗

Whether this whl package can be used to run the model on AMD or Hygon GPU(ROCm). Returns: Bool: `True` if ROCm is currently available, otherwise `False`. Examples: .. code-block:: pycon >>> import paddle >>> support_gpu = paddle.device.is_compi

()

Source from the content-addressed store, hash-verified

1012
1013
1014def is_compiled_with_rocm() -> bool:
1015 """
1016 Whether this whl package can be used to run the model on AMD or Hygon GPU(ROCm).
1017
1018 Returns:
1019 Bool: `True` if ROCm is currently available, otherwise `False`.
1020
1021 Examples:
1022 .. code-block:: pycon
1023
1024 >>> import paddle
1025 >>> support_gpu = paddle.device.is_compiled_with_rocm()
1026 """
1027 return core.is_compiled_with_rocm()
1028
1029
1030def cuda_places(

Callers 4

__init__Method · 0.85
affine_gridFunction · 0.85
conv2dFunction · 0.85
graphs.pyFile · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected