[PATCH v2 0/5] Improve fmod and fmodf

[Adhemerval Zanella]

This is an updated version of a previous submission aimed to improve
fmod implementation [1] by Kirill Okhotnikov.  I extended it with:

  1. Proper benchmarks for both single and double.  The inputs are
     divided in 3 subsets: subnormals, normal nubmers, and close 
     exponents.  It uses a list with random generated values.

  2. Use math_config.h definitions instead math_private (so it might
     eventually get back on optimize-routines).

  3. Implement the same strategy for float version.

  4. Also tuned the final division to use multiplication with inverse
     instead of direct modulo.  It showed better performance on both
     x86_64 and aarch64 chips I have tested.

  5. Remove SVID error handling wrapper.

The performance shows a good improvement compared to current algorithm
for fmod (using gcc 11):

  Architecture     | Input           | master   | patch
  -----------------|-----------------|----------|--------
  x86_64 (Ryzen 9) | subnormals      | 19.1584  | 9.40992
  x86_64 (Ryzen 9) | normal          | 1016.51  | 296.738
  x86_64 (Ryzen 9) | close-exponents | 18.4428  | 13.119
  aarch64 (N1)     | subnormal       | 11.153   | 4.33313
  aarch64 (N1)     | normal          | 528.649  | 158.339
  aarch64 (N1)     | close-exponents | 11.4517  | 5.76138

I also see similar improvements on arm-linux-gnueabihf when running on
the N1 aarch64 chips, where it uses a lot of soft-fp implementation
(for modulo, clz, ctz, and multiplication):

  Architecture     | Input           | master   | patch
  -----------------|-----------------|----------|--------
  armhf (N1)       | subnormal       | 15.7284  | 14.5746
  armhf (N1)       | normal          | 837.525  | 241.738
  armhf (N1)       | close-exponents | 16.2111  | 22.457


The fmodf shows a more moderate improvement:

  Architecture     | Input           | master   | patch
  -----------------|-----------------|----------|--------
  x86_64 (Ryzen 9) | subnormals      | 17.2549  | 9.35776
  x86_64 (Ryzen 9) | normal          | 85.4096  | 46.2761
  x86_64 (Ryzen 9) | close-exponents | 19.1072  | 12.6199
  aarch64 (N1)     | subnormal       | 10.2182  | 4.39188
  aarch64 (N1)     | normal          | 60.0616  | 18.3888
  aarch64 (N1)     | close-exponents | 11.5256  | 5.93518
  armhf (N1)       | subnormal       | 11.6662  | 7.75977
  armhf (N1)       | normal          | 69.2759  | 31.623
  armhf (N1)       | close-exponents | 13.6472  | 15.6689


I also checked against H.J proposal to use fprem on x86_64 [2] and
against recent suggestion on libc-alpha [3], and on both cases 
this newer implementation shows better performance.

Changes from v1:
 * Remove SVID error handling wrapper.
 * Extend testing for subnormal with different signs.
 * Code cleanup.

[1] https://sourceware.org/pipermail/libc-alpha/2020-November/119794.html
[2] https://patchwork.sourceware.org/project/glibc/patch/20230309183312.205763-1-hjl.tools@gmail.com/
[3] https://sourceware.org/pipermail/libc-alpha/2023-March/146164.html

Adhemerval Zanella (5):
  benchtests: Add fmod benchmark
  benchtests: Add fmodf benchmark
  math: Improve fmod
  math: Improve fmodf
  math: Remove the error handling wrapper from fmod and fmodf

 benchtests/Makefile                           |    2 +
 benchtests/fmod-inputs                        | 2182 +++++++++++++++++
 benchtests/fmodf-inputs                       | 2182 +++++++++++++++++
 math/Versions                                 |    4 +
 math/libm-test-fmod.inc                       |   11 +
 math/w_fmod_compat.c                          |   13 +-
 math/w_fmodf_compat.c                         |    6 +-
 sysdeps/i386/fpu/w_fmod_compat.c              |   14 +
 sysdeps/i386/fpu/w_fmodf_compat.c             |   14 +
 sysdeps/ieee754/dbl-64/e_fmod.c               |  245 +-
 sysdeps/ieee754/dbl-64/math_config.h          |   70 +
 sysdeps/ieee754/dbl-64/math_err.c             |    6 +
 sysdeps/ieee754/dbl-64/w_fmod.c               |    1 +
 sysdeps/ieee754/flt-32/e_fmodf.c              |  241 +-
 sysdeps/ieee754/flt-32/math_config.h          |   48 +
 sysdeps/ieee754/flt-32/math_errf.c            |    6 +
 sysdeps/ieee754/flt-32/w_fmodf.c              |    1 +
 sysdeps/m68k/m680x0/fpu/w_fmod_compat.c       |   14 +
 sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c      |   14 +
 sysdeps/unix/sysv/linux/aarch64/libm.abilist  |    2 +
 sysdeps/unix/sysv/linux/alpha/libm.abilist    |    2 +
 sysdeps/unix/sysv/linux/arm/be/libm.abilist   |    2 +
 sysdeps/unix/sysv/linux/arm/le/libm.abilist   |    2 +
 sysdeps/unix/sysv/linux/hppa/libm.abilist     |    2 +
 .../sysv/linux/m68k/coldfire/libm.abilist     |    2 +
 .../sysv/linux/microblaze/be/libm.abilist     |    2 +
 .../sysv/linux/microblaze/le/libm.abilist     |    2 +
 .../unix/sysv/linux/mips/mips32/libm.abilist  |    2 +
 .../unix/sysv/linux/mips/mips64/libm.abilist  |    2 +
 sysdeps/unix/sysv/linux/nios2/libm.abilist    |    2 +
 .../linux/powerpc/powerpc32/fpu/libm.abilist  |    2 +
 .../powerpc/powerpc32/nofpu/libm.abilist      |    2 +
 .../linux/powerpc/powerpc64/be/libm.abilist   |    2 +
 .../linux/powerpc/powerpc64/le/libm.abilist   |    2 +
 .../unix/sysv/linux/s390/s390-32/libm.abilist |    2 +
 .../unix/sysv/linux/s390/s390-64/libm.abilist |    2 +
 sysdeps/unix/sysv/linux/sh/be/libm.abilist    |    2 +
 sysdeps/unix/sysv/linux/sh/le/libm.abilist    |    2 +
 .../sysv/linux/sparc/sparc32/libm.abilist     |    2 +
 .../sysv/linux/sparc/sparc64/libm.abilist     |    2 +
 .../unix/sysv/linux/x86_64/64/libm.abilist    |    2 +
 .../unix/sysv/linux/x86_64/x32/libm.abilist   |    2 +
 42 files changed, 4923 insertions(+), 197 deletions(-)
 create mode 100644 benchtests/fmod-inputs
 create mode 100644 benchtests/fmodf-inputs
 create mode 100644 sysdeps/i386/fpu/w_fmod_compat.c
 create mode 100644 sysdeps/i386/fpu/w_fmodf_compat.c
 create mode 100644 sysdeps/ieee754/dbl-64/w_fmod.c
 create mode 100644 sysdeps/ieee754/flt-32/w_fmodf.c
 create mode 100644 sysdeps/m68k/m680x0/fpu/w_fmod_compat.c
 create mode 100644 sysdeps/m68k/m680x0/fpu/w_fmodf_compat.c

-- 
2.34.1