par2cmdline-turbo aims to keep close with its upstream and only differs in the following areas related to performance. As such, the scope of this project covers:
- GF16, MD5 and CRC32 computation, using ParPar’s implementation
- Use ParPar’s internal checksumming to detect RAM errors during GF16 computation
- Adopts ParPar’s stitched hashing approach
Note that this fork isn’t intended to change too much of par2cmdline itself and hence, is not aimed at fixing bugs or functionality improvements. Bug reports and feature requests should be directed at the upstream project where relevant. The minimalistic nature of this project also means that there’s some degree of performance left on the table, and there’s no focus on improving PAR1 performance.
Note that par2cmdline-turbo may use slightly more memory than par2cmdline, due to increased buffering in ParPar’s backend.
Your gratitude and finance will help me to continue integration of this QPKG and maintain up to date versions.