OpenJPH 0.28 Up To 1.9x Faster With New AVX2 Optimizations For High-Throughput JPEG2000
OpenJPH as an open-source implementation of high-throughput JPEG2000 Part-15 (or JPH or HTJ2K) is now significantly faster for both encode and decode operations thanks to new AVX2 optimizations for Intel and AMD processors.
This high-throughput JPEG2000 implementation is now even more high-throughput with new Advanced Vector Extensions 2 (AVX2) optimizations benefiting Intel and AMD x86_64 CPUs from the past number of years. OpenJPH 0.28 significantly improves encode and decode with new AVX2 block coder optimizations contributed by Prof Osamu Watanabe.
Here is the exciting takeaway from the new version:
More details on the new AVX2 optimizations can be found via the pull request.
Those making use of HTJ2K can find the new OpenJPH release via GitHub.
This high-throughput JPEG2000 implementation is now even more high-throughput with new Advanced Vector Extensions 2 (AVX2) optimizations benefiting Intel and AMD x86_64 CPUs from the past number of years. OpenJPH 0.28 significantly improves encode and decode with new AVX2 block coder optimizations contributed by Prof Osamu Watanabe.
Here is the exciting takeaway from the new version:
More details on the new AVX2 optimizations can be found via the pull request.
Those making use of HTJ2K can find the new OpenJPH release via GitHub.