Debian Patches
Status for hipfft/7.2.4-1~exp1
| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| include-README.md-doxygen.patch | Add README.md to Doxygen INPUT Doxygen is using USE_MDFILE_AS_MAINPAGE to a file that's outside of INPUT list. Marking as Forwarded: not-needed, I want a second opinion about his. |
Kari Pahula <kaol@debian.org> | not-needed | 2024-06-28 | ||
| fix-doxygen-refs.patch | fix doxygen refs reference errors when generating docs with sphinx via rocm-docs-build. Update comments to reference the correct public types and remove invalid struct annotations. No API or functional changes. |
Bojan Aleksovski <bojan.aleksovski@canonical.com> | yes | 2026-01-20 | ||
| use-local-mathjax.patch | use local mathjax The sphinx.ext.mathjax extension defaults to loading mathjax from a CDN, which results in the lintian warning 'privacy-breach-generic'. Use a local copy of mathjax to prevent that problem. |
Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2023-05-28 | ||
| Extend-docs-conf.py-for-offline-build.patch | Extend docs/conf.py for offline build By setting these extra variables, we can suppress a remote call which would cause the build to fail. |
Christian Kastner <ckk@debian.org> | no | 2025-06-29 | ||
| Add-version-script-to-control-exposed-symbols.patch | Add version script to control exposed symbols The C++ symbols emitted are clearly not intended to be part of the API. Rather than spending a lot of time finding the root cause, simply allow only the rocsparse_ symbols. Based on a solution by Xuanteng Huang for src:rocm-smi-lib. |
Christian Kastner <ckk@debian.org> | not-needed | 2025-06-29 | ||
| bump-rocm-docs-core-in-docs-sphinx.patch | Bump rocm-docs-core from 1.17.0 to 1.31.1 in /docs/sphinx Bumps [rocm-docs-core](https://github.com/ROCm/rocm-docs-core) from 1.17.0 to 1.31.1. - [Release notes](https://github.com/ROCm/rocm-docs-core/releases) - [Changelog](https://github.com/ROCm/rocm-docs-core/blob/develop/CHANGELOG.md) - [Commits](https://github.com/ROCm/rocm-docs-core/compare/v1.17.0...v1.31.1) |
"dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com> | yes | upstream | upstream, https://github.com/ROCm/hipFFT/pull/182/commits/da6734d12da1c1e57668941bbf998826cb475416.patch | 2026-01-01 |
| add-so-version-for-libhipfftw0.patch | Add versioning for hipfftw.so dfa828c45abcbeb804876b973314cf39d2b3b2e8 and in addition to being refreshed has a fix in the checking that hipfftw is not an alias: `if (aliased_hipfftw_name STREQUAL "")` is replaced with `if (NOT aliased_hipfftw_name)` because such if statement does not work when the variable is unset as in our case. This fix has been merged upstream (PR #4403) into the development branch; it is not part of any tagged release yet, so the patch is still carried. |
Bojan Aleksovski <bojan.aleksovski@canonical.com> | yes | upstream | upstream, https://github.com/ROCm/rocm-libraries/commit/dfa828c45abcbeb804876b973314cf39d2b3b2e8.patch | 2026-06-05 |
| use-free-vram-for-device-memory-check.patch | Use free VRAM for the device memory budget check accuracy_test.h uses hipMemGetInfo() to determine the available VRAM budget for test cases, but compares problem sizes against the TOTAL device memory rather than the FREE memory. On test machines where the GPU is already partially loaded (for example, when the hipfft accuracy suite runs alongside other ROCm autopkgtests on a shared GPU), this causes large tests to pass the skip check, allocate their main buffers, and then run the FFT with no room left for rocFFT work buffers -- producing large relative errors instead of being skipped. . Using free VRAM ensures tests that cannot fit in the currently available device memory are skipped rather than producing spurious accuracy failures. |
Igor Luppi <igor.luppi@canonical.com> | not-needed | 2026-06-04 |
All known versions for source package 'hipfft'
- 7.2.4-1~exp1 (experimental)
- 6.4.3-2 (sid)
- 6.1.2-3 (trixie)
