Skip to content

Latest commit

 

History

History
114 lines (78 loc) · 3.15 KB

CHANGELOG.md

File metadata and controls

114 lines (78 loc) · 3.15 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

Added

  • Added benchmarks for glam's affine types

Changed

  • Updated euclid to 0.22.7
  • Updated glam to 0.21.3
  • Updated nalgebra to 0.31.1
  • Updated static-math to 0.2.3
  • Updated ultraviolet to 0.9.0
  • Updated vek to 0.15.8

0.4.0 - 2020-11-10

Added

  • Added wide benchmarks for nalgebra and ultraviolet libraries.

Changed

  • Updated glam to 0.10.0
  • Updated nalgebra to 0.23.0
  • Updated static-math to 0.1.7
  • Updated ultraviolet to 0.7.1

0.3.2 - 2020-09-05

Added

  • Added the static-math library.
  • Added scalar benchmarks for the ultraviolet library.
  • Added section on wide benchmarks.

Changed

  • Updated glam to 0.9.4 and changed benchmarks to use Vec3A instead of Vec3
  • Updated nalgebra to 0.22.0
  • Updated euclid to 0.22.1
  • Updated vek to 0.12.0
  • Updated pathfinder_geometry to 0.5.1
  • Updated ultraviolet to 0.5.1 (currently build bench only)

0.3.1 - 2020-04-10

Added

  • Added buildbench tool for comparing build times of the math libraries supported by mathbench.

0.3.0 - 2019-12-03

Added

  • Added "return self" benchmarks to try establish the overhead of running a benchmark for each type.

Changed

  • Updated glam to 0.8.3 and euclid to 0.20.5.
  • Split 2D and 3D transform types out of 3x3 and 4x4 matrix benches into their own specialised transform benches.
  • Run some benches over different input sizes, mostly 1 and 100 for binary operations, to get a better idea of a single operation versus operating over an array of data.
  • Updated summary script to deal with input sizes and to highlight all benchmark times that are within a configurable threshold of the minimum time.

0.2.0 - 2019-11-25

Added

  • Added benchmarks for pathfinder_geometry 0.3.0

Changed

  • Changed benchmarks to store output values instead of discarding them.
  • Updated euclid to 0.20.4

0.1.9 - 2019-11-13

Added

  • Added this CHANGELOG.md file!
  • Added benchmarks for vek 0.9.10

Changed

  • cgmath, euclid and nalgebra are now all optional features. They can be excluded from benchmarks using cargo bench --no-default-features.
  • Updated glam to 0.8.2
  • Updated nalgebra to 0.19.0
  • Updated euclid to 0.20.3
  • Updated criterion to 0.3.0
  • Benchmarks now use criterion's BenchmarkGroup struct. The old way of benchmarking functions has been silently deprecated in 0.3.0.