Minimal-Precision Computing for High-Performance, Energy-Efficient, and Reliable Computations - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2020

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

Résumé

We have recently started a research collaboration to explore the possibility of a new computing system with precision-tuning, in collaboration with RIKEN CCS, Sorbonne University, and University of Tsukuba. Our proposal, "minimal-precision computing," aims to achieve reliability (accuracy and reproducibility) as well as high-performance (speed and energy) by obtaining the computing results with the accuracy requested by users with the minimal-precision use. Our proposal involving both hardware and software stacks combines (1) a precision-tuning method through numerical validation by Discrete Stochastic Arithmetic (DSA), (2) arbitrary-precision arithmetic libraries, (3) fast and accurate numerical libraries, and (4) Field-Programmable Gate Array (FPGA) with high-level synthesis, and some important components that we develop. This poster introduces an overview of our ideas and our up-to-date contributions.
Fichier non déposé

Dates et versions

hal-02486850 , version 1 (21-02-2020)

Identifiants

  • HAL Id : hal-02486850 , version 1

Citer

Taisuke Boku, Norihisa Fujita, Stef Graillat, Jens Huthmann, Roman Iakymchuk, et al.. Minimal-Precision Computing for High-Performance, Energy-Efficient, and Reliable Computations. 2nd R-CCS international symposium, K to Fugaku: Simulation, Big Data and AI supporting Society 5.0, Feb 2020, Kobe, Japan. ⟨hal-02486850⟩
166 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More