Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

GPU: Add Device::deviceVendor() function and fix #3416 #3418

Merged
merged 2 commits into from
Jul 12, 2023
Merged

GPU: Add Device::deviceVendor() function and fix #3416 #3418

merged 2 commits into from
Jul 12, 2023

Commits on Jul 12, 2023

  1. GPU: Add Device::deviceVendor() function 

    Note that this is only really relevant for SYCL. For CUDA and HIP is is
inferred at compile time.
    @mirenradia
    mirenradia committed Jul 12, 2023
    Configuration menu
    Copy the full SHA
    ea99ddd View commit details
    Browse the repository at this point in the history

  2. SYCL: Default disable SIGSEGV handling on Intel GPUs 

    SIGSEGV handling currently interferes with the managed memory
implementation on Intel's discrete GPUs. In theory, we could only
disable sigsegv handling for specific Intel GPU architectures (e.g.
using the experimental oneAPI SYCL Device Architecture extension
(https://github.com/intel/llvm/blob/48be219e238ca2354a1e9e9989752aa7a60b5627/sycl/doc/extensions/experimental/sycl_ext_oneapi_device_architecture.asciidoc)
but this would need to be updated every time Intel releases a new GPU
architecture which still uses SIGSEGV for managed memory. It is simpler
to disable it by default on all Intel GPUs.

This fixes AMReX-Codes#3416.
    @mirenradia
    mirenradia committed Jul 12, 2023
    Configuration menu
    Copy the full SHA
    4149659 View commit details
    Browse the repository at this point in the history