Skip to content
This repository was archived by the owner on Mar 21, 2024. It is now read-only.

Milestones

List view

  • Last planned release of the 1.X line. First RC planned for May 9.

    Due by May 23, 2022
    17/17 issues closed
  • First RC planned for 2/9.

    Due by February 14, 2022
    29/29 issues closed
  • First RC planned for 10/18.

    Due by November 1, 2021
    17/17 issues closed
  • Included with NVHPC 21.9. First RC planned for 8/18.

    Due by August 25, 2021
    14/14 issues closed
  • Major release. Included in NVHPC 21.7. Feature freeze / first RC planned for 6/14.

    Due by June 28, 2021
    21/21 issues closed
  • Major release. Included in NVHPC 21.3. Expands collection of `thrust::async` algorithms to include `inclusive_scan`/`exclusive_scan`. New `thrust::universal_vector<T>` to provide convenient access to CUDA unified memory. Deprecate Clang < 7, MSVC < 2019. First RC planned for 2/21.

    Due by February 28, 2021
    36/36 issues closed
  • Improved sort performance for CUDA backend. New variadic APIs. Various bugfixes, modernizations, and performance improvements.

    Due by November 30, 2020
    37/37 issues closed
  • Minor release. We will use a release branch, `staging/1.9.10-1`, based at the `1.9.10` tag, for this release.

    Due by July 10, 2020
    20/20 issues closed
  • Due by May 1, 2020
    16/16 issues closed
  • Major release. Included in NVHPC 20.9. Drops support for C++03. Overhauls CMake support.

    Due by September 15, 2020
    32/32 issues closed
  • Deprecation of C++03/C++11 and GCC <5. Support for MSVC <2015 dropped.

    Due by April 23, 2020
    8/8 issues closed
  • CUB added as a submodule and made a first class citizen in the CUDA Toolkit. Fix 32-bit indexing bug in all algorithms except sort.

    Due by February 21, 2020