GPU problem with Matlab 2017b

1 visualización (últimos 30 días)
caesar
caesar el 1 de Dic. de 2017
Comentada: caesar el 4 de Dic. de 2017
I have two Nvidia Tesla C 1060 connected to the machine (workstation)that runs Matlab 2017b. I ran this command :
gpuDevice
to check the compatibility. I got this error message :
Error using gpuDevice (line 26)
There is a problem with the graphics driver or with this GPU device. Be sure that you have a
supported GPU and that the latest driver is installed.
Caused by:
The graphics driver was found, but it is too old. The graphics driver on your system supports CUDA version 6.5. The required CUDA version is 8 or greater.
tried to install the recent CUDA release from Nvidia (cuda_9.0.176_windows_network) and it didn't work.
installed the recent driver for the GPU card and CUDA 6.5 (cuda_6.5.14_windows_general_64) and no change at all.

Respuestas (1)

Joss Knight
Joss Knight el 3 de Dic. de 2017
Editada: Joss Knight el 3 de Dic. de 2017
To use a Fermi card with R2017b you will first need to install Update 2. Follow the link in this bug report: https://www.mathworks.com/support/bugreports/1668715
Secondly, NVIDIA have some backwards compatibility issues with Fermi cards and the latest drivers. You should install a CUDA 8 driver from the 37x series, such as this one: http://www.nvidia.com/download/driverResults.aspx/113456/en-us
It's worth noting that NVIDIA are no longer posting any certified drivers for the C1060 on their website, so it may well be that they have fully deprecated that card, and there could still be some other issues. However, drivers for any compatible card are normally fine. If there were other problems your only solution would be to not upgrade MATLAB and keep your original driver.
  3 comentarios
Edric Ellis
Edric Ellis el 4 de Dic. de 2017
Support for the C1060 device was removed from Parallel Computing Toolbox in R2014b, see this release note.
caesar
caesar el 4 de Dic. de 2017
Thanks Edric Ellis, now its all clear.

Iniciar sesión para comentar.

Categorías

Más información sobre GPU Computing en Help Center y File Exchange.

Community Treasure Hunt

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

Start Hunting!

Translated by