tensorflow-rocm needs hipErrorNoBinaryForGpu (does it need recompile?) about rocm-build HOT 7 CLOSED

Yes, you need recompile tensorflow with gfx1030, but itis no guanrontee that we can run ROCm on gfx1030 properly.

Good news is ROCm said they will support navi in 2021. So my suggestion is waiting for official support.

from rocm-build.

I guess gfx1030 will be supported on ROCm-4.3.
https://github.com/ROCmSoftwarePlatform/rocBLAS/blob/release/rocm-rel-4.3/CMakeLists.txt#L169

ROCm-4.3 may be released about June or July 2021, I wish.

from rocm-build.

Indeed. Thanks for this project.

from rocm-build.

Hi.

I have ROCm 4.3 with gfx1030. I installed tensorflow-rocm and when it imports fine. But when I run to check if it's using GPU, then I get this:

tf.config.list_physical_devices('GPU')

2021-09-11 09:15:31.830208: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libamdhip64.so
"hipErrorNoBinaryForGpu: Unable to find code object for all current devices!"
Aborted (core dumped)

Any suggestions?

from rocm-build.

Sadly ROCM still doesnt supper Navi. I was also hoping 4.3 would. Some of the libraries did add support.

from rocm-build.

Which deep learning frameworks can I use then?

from rocm-build.

@Flock1
The tensorflow cost more time than pytorch.
If I have a gfx1030, I will recompile pytorch to test.

Here is some scipts for gfx1030:
https://github.com/xuhuisheng/rocm-build/tree/master/navi21

from rocm-build.