Skip to content
  1. Oct 28, 2024
  2. Oct 24, 2024
    • Pierre Blanchard's avatar
      Update GH runners to Ubuntu 24.04 · 6abb7d92
      Pierre Blanchard authored
      This will ultimately unlock the use of latest versions of
      gcc 12, then gcc 13, and finally gcc 14.
      Gcc 14 will unlock more testing.
      This PR does not update version of gcc yet, just to make sure
      old version still pass tests.
      Remove some expensive tests.
      6abb7d92
  3. Oct 18, 2024
  4. Oct 16, 2024
    • Naoki Shibata's avatar
      This patch fixes two problems (#592) · f95a6c86
      Naoki Shibata authored
      
      
      This patch fixes the following two problems:
      
      * On aarch32, cbrt(4) was returned instead of M_PI/2 when +-INFINITY is given to atanf.
      * Due to a bug in gcc-13, sign of tanf(+-0) was wrong on PUREC_SCALAR and VXE.
      
      This patch also adds testing for the first problem.
      The second problem is detected by the existing tester.
      
      ---------
      
      Co-authored-by: default avatarNaoki Shibata <shibatch.sf.net@gmail.com>
      f95a6c86
  5. Oct 15, 2024
    • Joana Cruz's avatar
      Integrate Google benchmarks into SLEEF · bbdf0e2d
      Joana Cruz authored
      Added new benchmark tool to SLEEF project using googlebench
      framework.
      In theory this tool can benchmark any unary and binary function
      in SLEEF.
      Benchmark is enabled in all functions listed in benchsleef.cpp.
      This list uses macros in benchmark_callers.cpp, so that we can
      enable benchmarking in multiple precisions using single lines
      of code. It is also possible to list the exact function we want,
      as each of the macros in benchmark_callers.cpp can be called
      independently and connected like building blocks.
      The tool is integrated with SLEEF via CMake, meaning it can
      be built automatically when SLEEF is built. To enable that,
      pass CMake argument -DSLEEF_BUILD_BENCH=ON.
      
      This tool depends on C++17 standard.
      Tested on aarch64 for scalar, vector and SVE routines
      Tested on x86 for different vector length extensions.
      Tested for llvm-17, gcc-11 and gcc-14.
      bbdf0e2d
  6. Oct 14, 2024
  7. Oct 10, 2024
  8. Oct 09, 2024
    • Pierre Blanchard's avatar
      Add checks for incompatible changes in GHA · 695e734c
      Pierre Blanchard authored
      Add a new workflow to detect unexpected backward-incompatible
      changes at precommit, by comparing current version against a
      reference version, both generated on Linux with gcc.
      It uses abidiff to compare ELF-format shared libraries.
      This is the tool used by Fedora to report a recent breach in
      API #534. Test should fail until it is fixed by #545.
      695e734c
  9. Oct 04, 2024
    • Pierre Blanchard's avatar
      Use Arm-hosted runners (#581) · 8373ab19
      Pierre Blanchard authored
      Documentation for these runners here https://gitlab.arm.com/tooling/gha-runner-docs
      Note that only SLEEF upstream repo is allowed to use these runners.
      
      Integrate runners via strategy.matrix.os, in order to facilitate customisation
      of step (e.g. config) based on runner label.
      
      Install OpenSSL since tester3 is enforced.
      New runners do not seem to come with openssl or cmake pre-installed.
      
      Use SVE enabled instances (c7g)
      Use 2xlarge instances so tests run slightly faster without compromising on availability.
      
      Remove aarch64 from cross build and tests (on Linux).
      And update documentation/README.
      8373ab19
  10. Oct 02, 2024
  11. Sep 17, 2024
  12. Sep 16, 2024
  13. Sep 09, 2024
    • Joana Cruz's avatar
      Update documents for acosh · 68f5497a
      Joana Cruz authored
      The acosh function is only defined for values greater or
      equal to 1. Documentation was updated according to this.
      Also, rephrased a confusing sentence.
      68f5497a
  14. Aug 29, 2024
  15. Aug 12, 2024
  16. Jul 26, 2024
  17. Jul 17, 2024
    • Joe Ramsay's avatar
      Enable GNUABI build on more targets #525 (#562) · 0e47fcc3
      Joe Ramsay authored
      * Enable GNUABI build on more targets #525
      
      Previously SLEEF_BUILD_GNUABI_LIBS was silently ignored except for a
      small number of targets, or if the compiler does not support weak
      aliases.
      
      It can now be built regardless of OS and compiler on AArch64 and
      x86_64 - trying to enable it on any other target is now an error.
      
      The GLIBC *_finite symbols are handled using the same workaround as is
      used for the DALIAS macro when compiling on a target which does not
      support aliases. Though in practice these symbols are unlikely to be
      required on systems where aliases are unsupported, they are part of
      the API so need adding, if just to make the tests build.
      
      Also removed two CMake variables:
      - ENABLE_ALIAS, which was unused
      - ENABLE_GNUABI, which was now redundant but could still be forced on,
        leading to cryptic build failures
      Both names are still used by the preprocessor in sleefsimd* sources
      for managing names and aliases.
      
      * Fix masked GNUABI build for NOALIAS
      
      Masked symbols also need to sidestep aliases when they are not
      supported.
      
      * Disable GNUABI for unsupported targets in precommit
      
      Default is on, but this setting was previously just ignored when
      unsupported. Disable it in the failing pipelines.
      0e47fcc3
  18. Jun 10, 2024
  19. Jun 07, 2024
Loading