Created attachment 309131 [details, diff] Patch to current stable for 27.4 version Version 27.4 of sci-mathematics/gimps is available for download. This is a beta for the new version. This version supports 64-bit optimized AVX FFTs. 32-bit AVX FFTs are also a little bit faster. This version fixes a number of bugs in 27.3 (which I never requested to be added to the tree in the first place, too beta) - mostly related to LLR and PFGW issues. Unless one has a Sandy Bridge (or later) Intel CPU, there is little reason to upgrade to this version. Known bugs and fixes: 1. Does not work on AMD Bulldozer processors. 2. Length 1280K FFT crashes on Pentium 4s with 512K L2 cache or less. 3. Hyperthread detection is not working properly. 4. The 6K 2^N-1 and 8K, 10K, 12K, 16K 2^N+1 AVX FFTs are missing. 5. Some exponents above 536 million generate roundoff errors and incorrect results. 6. FFT crossover points need retuning. Many can be more aggressive, but a few need to be more conservative. This is not yet a candidate for stable, but it would be nice to have it in ~arch (or even package.mask'ed) for beta-testing. Versions below 26.6 should be removed from the tree.
Hi, > > This is not yet a candidate for stable, but it would be nice to have it in > ~arch (or even package.mask'ed) for beta-testing. Versions below 26.6 should > be removed from the tree. I guess ~arch is fine. Gimps users know what they are doing ;) I added it. I also migrated to EAPI=4 and removed old ebuilds. Thanks a lot!
