MATLAB R2019b does not use CUDA 10.1

4 次查看(过去 30 天)
Matlabs seems not to use the correct CUDA driver and throws an error.
MatlabR2019b running on Ubuntu with all the required CUDA toolkits installed throws this error:
>> gpuDevice
Error using gpuDevice (line 26)
CUDA driver version 10 is not supported. Update CUDA driver to version 10.1 or greater. For more information
on GPU support, see GPU Support by Release.
Anyone an idea on how to make Matlab use the cuda-10.1?

采纳的回答

Jason Ross
Jason Ross 2019-10-2
Run nvidia-smi in a terminal window and see what driver version is reported. It sounds like you have a 10.0 driver and need to update it to 10.1.

更多回答(2 个)

cui,xingxing
cui,xingxing 2019-10-9
Matlab can be compatible with the previous cuda version, or upgrade the matlab every time you need to upgrade the cuda version, I feel so troublesome. . .

Mohammad Hossein Rafiei
I am having a similar issue. I have a NVIDIA GeForceGTX 1070. Upgraded to version 10.1 and it still not working. Matlab errors out "CUDA driver version 9.1 is not supported. Update CUDA driver to version 10.1 or greater. For more information on GPU support, see GPU Support by Release." Any thoughts?
  1 个评论
Jason Ross
Jason Ross 2019-10-16
Did you update the toolkit and SDK to 10.1 and not the driver? What ultimately matters is the driver. You should be able to see that when you run nvidia-smi in the upper right hand corner of the output. On Linux it's generally on the path, on Windows it's in C:\Program Files\Nvidia corporation\nvsmi\nvidia-smi.exe
Generally speaking, the driver should always be at the latest version, as nVidia provides compatibility for earlier versions in newer drivers. It doesn't work the other way around -- older drivers by definition don't have the required libraries to support newer toolkits.

请先登录,再进行评论。

类别

Help CenterFile Exchange 中查找有关 GPU Computing 的更多信息

产品


版本

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by