Poster 134: Minimal-Precision Computing for High-Performance, Energy-Efficient, and Reliable Computations

Authors: Daichi Mukunoki (RIKEN Center for Computational Science (R-CCS)), Toshiyuki Imamura (RIKEN Center for Computational Science (R-CCS)), Yiyu Tan (RIKEN Center for Computational Science (R-CCS)), Atsushi Koshiba (RIKEN Center for Computational Science (R-CCS)), Jens Huthmann (RIKEN Center for Computational Science (R-CCS)), Kentaro Sano (RIKEN Center for Computational Science (R-CCS)), Fabienne Jézéquel (Sorbonne University), Stef Graillat (Sorbonne University), Roman Iakymchuk (Sorbonne University), Norihisa Fujita (University of Tsukuba), Taisuke Boku (University of Tsukuba)

Abstract: In numerical computations, the precision of floating-point computations is a key factor to determine the performance (speed and energy-efficiency) as well as the reliability (accuracy and reproducibility). However, the precision generally plays a contrary role for both. Therefore, the ultimate concept for maximizing both at the same time is the minimal-precision computation through precision-tuning, which adjusts the optimal precision for each operation and data. Several studies have been already conducted for it so far, but the scope of those studies is limited to the precision-tuning alone. In this study, we propose a more broad concept of the minimal-precision computing with precision-tuning, involving both hardware and software stack.

Best Poster Finalist (BP): no

Poster: PDF
Poster summary: PDF


Back to Poster Archive Listing