quantum-computingqristal-sdk

Is there a way to force Qristal SDK to use available NVIDIA GPUs by default?

I've been using the 1.0 release of the Qristal SDK via the Docker image, and have successfully run some circuits. I'm not sure if it is using the available GPUs however, and I'm not sure how to check this. Can I force it at runtime, or control in the running session?

Solution

  • If your system has one or more NVIDIA GPUs, install the NVIDIA Container Toolkit and add the switch --gpus all to the docker run command in order to use them.

    So this would look like:

    docker run --rm -it --name --gpus all qbsdk -d -p 8889:8889 registry.gitlab.com/qbau/software-and-apps/public/qbsdk

    Referencing the updated the quickstart guide. Good catch.