General Mechanical

General Mechanical

How to enable GPU acceleration properly on ANSYS Mechanical Enterprise?

    • ashwin11
      Subscriber

      Hi All

      I tried enabling GPU acceleration on ANSYS Mechanical but I am not sure whether the GPU is being utilized in the processing. I have done the following as of now

      1) Set the ANSGPU_OVERRIDE variable to 1.

      2) Enabled the number of GPU to be used in the solver settings to 2.

      I still get the following on the solver output file

      "The Distributed Sparse Matrix Solver was not able to sufficiently utilize the GPU accelerator device(s) during this equation solution. This can be caused by one or more factors such as the size of the system of equations being solved; using older or less powerful GPU hardware relative to the CPU core(s) involved; having insufficient available memory on the GPU device. The check for GPU performance terminates for the remainder of this session."

      My Machine is a Dell Precision 7920 Tower with 48 physical cores, 256 GB RAM and 2*16 GB GPU NVIDIA Quadro RTX 5000.

      Any help would be appreciated!

      Thank you

      Ashwin

    • Gary Stofan
      Ansys Employee

      The Quadro RTX5000 is not supported. As the Help states, you can often force the GPU to be recognized with the override, but other logic prevents it from being used.
      I suggest obtaining a newer GPU that we list as supported. We recommend purchasing the card from a vendor that will accept product returns in case the new GPU doesn't provide the expected performance increase.
      https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v221/en/installation/win_gpureqt.html?q=gpu%20tesla
      To utilize a NVIDIA GPU device that is not on the recommended list of cards, set the following environment variable:
      ANSGPU_OVERRIDE=1
      This is most beneficial when you wish to run on newer NVIDIA GPUs that were not available at the time of release of this version of the Ansys program. If you choose to use this environment variable, you should ensure that the NVIDIA GPU device that you wish to use is sufficiently powerful, in terms of both double-precision compute power and on-card memory, to achieve meaningful acceleration for your simulation. Using this environment variable with an underpowered CPU may actually decelerate your simulation.
Viewing 1 reply thread
  • You must be logged in to reply to this topic.