Performance Tuning of Fock Matrix and Two-Electron Integral Calculations for NWChem on Leading HPC Platforms

Conference · Wed Oct 01 00:00:00 EDT 2014

DOI:https://doi.org/10.1007/978-3-319-10214-6_13· OSTI ID:1178519

Shan, Hongzhan; Austin, Brian M.; De Jong, Wibe A.; Oliker, Leonid; Wright, Nicholas J.; Apra, Edoardo

Attaining performance in the evaluation of two-electron repulsion integrals and constructing the Fock matrix is of considerable importance to the computational chemistry community. Due to its numerical complexity improving the performance behavior across a variety of leading supercomputing platforms is an increasing challenge due to the significant diversity in high-performance computing architectures. In this paper, we present our successful tuning methodology for these important numerical methods on the Cray XE6, the Cray XC30, the IBM BG/Q, as well as the Intel Xeon Phi. Our optimization schemes leverage key architectural features including vectorization and simultaneous multithreading, and results in speedups of up to 2.5x compared with the original implementation.

• Research Organization: Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 1178519
• Report Number(s): PNNL-SA-99759; KP1704020
• Resource Relation: Conference: 4th International Workshop on High Performance Computing Systems. Performance Modeling, Benchmarking and Simulation (PMBS13), November 18, 2013, Denver, Colorado. Lecture Notes in Computer Science, 8551:261-280
• Country of Publication: United States
• Language: English