A new data conversion method for mixed precision Krylov solvers with FP16/BF16 Jacobi preconditioners-Reference-Cited by-同舟云学术

A new data conversion method for mixed precision Krylov solvers with FP16/BF16 Jacobi preconditioners

Published:2023-02-27 Issue: Volume: Page:
ISSN:
Container-title:Proceedings of the International Conference on High Performance Computing in Asia-Pacific Region
language:
Short-container-title:

Author:

Ina Takuya¹^ORCID,Idomura Yasuhiro¹^ORCID,Imamura Toshiyuki²^ORCID,Onodera Naoyuki¹^ORCID

Affiliation:

1. Center for Computational Science and e-Systems, Japan Atomic Energy Agency, Japan

2. Center for Computational Science RIKEN, Japan

Publisher

ACM

Link

https://dl.acm.org/doi/pdf/10.1145/3578178.3578222

Reference17 articles.

1. 2016. PRIMERGY CX1640 M1. https://jcahpc.jp/eng/ofp_intro.html 2016. PRIMERGY CX1640 M1. https://jcahpc.jp/eng/ofp_intro.html

2. 2018. IBM Power System AC922. http://www.olcf.ornl.gov/olcf-resources/compute-systems/summit 2018. IBM Power System AC922. http://www.olcf.ornl.gov/olcf-resources/compute-systems/summit

3. 2020. HPE SGI 8600. https://ccse.jaea.go.jp/computer/index_eng.html 2020. HPE SGI 8600. https://ccse.jaea.go.jp/computer/index_eng.html

4. 2020. Supercomputer Fugaku. https://www.r-ccs.riken.jp/en/fugaku/about/ 2020. Supercomputer Fugaku. https://www.r-ccs.riken.jp/en/fugaku/about/

5. A. Haidar , S. Tomov , J. Dongarra , and N.J. Higham . 2018. Harnessing GPU Tensor Cores for Fast FP16 Arithmetic to Speed up Mixed-Precision Iterative Refinement Solvers . In SC18: International Conference for High Performance Computing, Networking, Storage and Analysis. 603–613 . A.Haidar, S.Tomov, J.Dongarra, and N.J. Higham. 2018. Harnessing GPU Tensor Cores for Fast FP16 Arithmetic to Speed up Mixed-Precision Iterative Refinement Solvers. In SC18: International Conference for High Performance Computing, Networking, Storage and Analysis. 603–613.