Skip to content

Releases: greg7mdp/sparsepp

2026.1 Stable

Choose a tag to compare

@greg7mdp greg7mdp released this 13 Jun 12:21

What's Changed

  • small fix for compiling on clang (Apple LLVM version 8.0.0 (clang-800.0.42.1)) by @jhetherly in #40
  • "gcc -Wextra" compilation warning mitigation by @Roslaniec in #94
  • Update conan support to conanv2. Add packaging test. by @studiofuga in #96
  • Fix compilation warnings with gcc-15. by @greg7mdp in #99
  • Fix warning on realloc call. by @greg7mdp in #100

New Contributors

Full Changelog: 1.22...1.23

2018.2 Stable

Choose a tag to compare

@greg7mdp greg7mdp released this 17 Sep 13:42

Fix issues #62 and #70

2018 Stable

Choose a tag to compare

@greg7mdp greg7mdp released this 19 Jun 21:46
1.21

fix warning (conversion from 'uint64_t' to 'std::size_t'), Issue #65,…

2017 Stable

Choose a tag to compare

@greg7mdp greg7mdp released this 08 Dec 21:48

Includes small changes in the last year. No code changes for a couple months. Calling this stable :-)

File reorganization - not single header anymore

Choose a tag to compare

@greg7mdp greg7mdp released this 21 Feb 02:50
  • not single header anymore - too much of a hassle
  • custom allocator not quite ready yet. It is checked in, but sparsepp is still using the old allocator (easy to toggle - line 15 of spp_config.h)

Stable release - main issue is the excessive memory usage on Windows

Choose a tag to compare

@greg7mdp greg7mdp released this 05 Feb 18:35

The first official release of sparsepp - production ready on linux/macos.

I do not consider it production ready on Windows, because the realloc implementation on this platform seems to waste an inordinate amount of memory when reallocating (and sparsepp does a lot of reallocations), so the memory usage is much greater on Windows than it should be.

To address this issue, I am currently implementing a custom allocator which will fix this memory usage issue on windows, and should also increase performance on all platforms. This will be version 1.0, as it will truly fulfill the promise of low memory usage on all platforms.