CA-BPN is a deterministic heuristic for choosing batch sizes in theta-coordinate isogeny computations, ensuring cache efficiency and microarchitectural scheduling. It solves the problem of optimizing batch sizes for constant-time normalization in isogeny pipelines. The project provides a reference implementation, sweep drivers, and sample measurements, making it useful for researchers and developers in the field of cryptography. CA-BPN stands out due to its focus on cache-aware batch size selection, which improves performance and efficiency.
View on GitHub →yui9696/sqisign-cabpn-artifact