Reproduce “Tenon: A Unified Programming Framework for Retargetable PIM Compilation” 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

Niansong Zhang, Shiran Guo, Zhiru Zhang

Tenon provides a shared compiler framework for mapping workloads and data across processing-in-memory systems. It describes compute and storage hierarchies, legal transfers, and costs, then uses binary linear layouts to search joint placements. On simulated UPMEM, Tenon is fastest on all 16 reported kernels, with a 1.20× geometric-mean speedup over the fastest baseline; backend code generation and runtimes remain device-specific.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.