Reproduce “Vectorizing Recurrences over Semirings” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yihong Zhang, Bala Vinaithirthan, Karima Ma, Zachary Tatlock, Maaz Bin Safeer Ahmad, Andrew Adams
VecRec rewrites fixed-order recurrences to expose parallel work for CPU vectorization, then generates and profiles equivalent schedules. Its dilation transformation widens feedback spacing for time-invariant kernels over rings or idempotent semirings. In Minimap2 exact mode on Zen 5, VecRec reduced ecoli-ultralong runtime from 58,235 ms to 45,624 ms, a 1.28× speedup; this result covers exact-mode workloads, not approximate modes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.