SC17 Denver, CO

P37: PaSTRI: A Novel Data Compression Algorithm for Two-Electron Integrals in Quantum Chemistry

Authors: Ali Murat Gok (Argonne National Laboratory, Northwestern University), Dingwen Tao (University of California, Riverside), Sheng Di (Argonne National Laboratory), Vladimir Mironov (Lomonosov Moscow State University), Yuri Alexeev (Argonne National Laboratory), Franck Cappello (Argonne National Laboratory)

Abstract: Integral computations for two-electron repulsion energies are very frequently used applications in quantum chemistry. Computational complexity, energy consumption and the size of the output data generated by these computations scales with O(N4), where N is the number of atoms simulated. Typically, the same integrals are calculated multiple times. Storing these values and reusing them requires impractical amounts of storage space; whereas recalculating them requires a lot of computations. We propose PaSTRI (Pattern Scaling for Two-electron Repulsion Integrals), a fast novel compression algorithm which makes it possible to calculate these integrals only once, store them, and reuse them at much smaller computational cost then recalculation. PaSTRI is “lossy” compared to floating point numbers, but still maintains the precision level required by the integral computations. PaSTRI is an extension to SZ compressor package as a part of ECP-EZ. PaSTRI achieves 17.5:1 compression ratio whereas vanilla SZ achieves 8.0:1 and ZFP achieves 7.1:1.
Award: Best Poster Finalist (BP): no

Poster: pdf
Two-page extended abstract: pdf

Poster Index