Release Notes:
- Improved performance of AVX-FMA, AVX-FMA-4, and AVX-512 iteration algorithms: 5% to 20%+ faster than QuickMAN 2.0, depending on CPU.
- Pan and zoom rates for logfile images now use the current slider values (instead of quickman.cfg values) if not explicitly set in logfile.
- 32-bit integers for configuration settings can now be specified as four individual bytes if desired.
- Fixed out of memory error if quickman.cfg was missing.
- Fixed slideshow stopping if duration elapsed before a zoom finished.
Example benchmarks, AVX-FMA-4 algorithm:
- i9-13900K: 1878 GFlops single precision / 933 GFlops double precision (+8.2% / +6.0% faster than QuickMAN 2.0)
- i7-5600U : 139.0 GFlops single precision / 69.5 GFlops double precision (+21.5% / +21.5% faster than QuickMAN 2.0)