Reproduce “SecDMP: Provably Secure Data Memory-Dependent Prefetching (Extended Version)” 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
Jens Sprengers, Marton Bognar, Lesly-Ann Daniel, Frank Piessens
SecDMP keeps data memory-dependent prefetchers (DMPs) from using secret data while retaining prefetching from public data. The authors prove it adds no leakage beyond a no-DMP baseline and evaluate it on simulated Proteus. In tests, it prevents DMP leakage where selective disabling leaves indirect cases exposed. Its average cycle-count overhead is 0.36% on DMP-Bench; these are simulations, not commercial-CPU measurements.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.