Performance additions surrounding log parsing#36

Merged

tdulcet merged 13 commits intotdulcet:devfrom

Apr 1, 2026

Commits on Mar 28, 2026

perf: replace /proc/cpuinfo regex with equivalent string parser
jcoxdco
committed
perf: hoist Mlucas stat and GpuOwl log regexes to module scope
jcoxdco
committed

Commits on Mar 31, 2026

remove verbose comments and inline single line function calls per feedback. Benchmarking the guarded lines code went from 3.03 to 4.27x faster.
jcoxdco
committed
Implement optimized line reading and benchmarking enhancements in autoprimenet.py, benchmarked 64, 256 and 1024 chunk sizes, 256 was most optimal
jcoxdco
committed
Simplified _iter_lines_reversed_chunked. Performace gain over prevous commit was 8-13%. Removed benchmark code and arguments.
jcoxdco
committed
removing cpuinfo benchmark after testing on WSL. OS reported as Ubuntu 22.04, kernel 6.6.87.2-microsoft-standard-WSL2, x86_64. Tests pass.
jcoxdco
committed
final tidy of comments to keep the diff small
jcoxdco
committed

Commits on Apr 1, 2026

Updated to simplify code.
tdulcet
committed