Enabling mixed-precision with the help of tools: A Nekbone case study
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909206775332864 |
|---|---|
| author | Chen, Yanxiang Castro, Pablo de Oliveira Bientinesi, Paolo Iakymchuk, Roman |
| author_facet | Chen, Yanxiang Castro, Pablo de Oliveira Bientinesi, Paolo Iakymchuk, Roman |
| contents | Mixed-precision computing has the potential to significantly reduce the cost of exascale computations, but determining when and how to implement it in programs can be challenging. In this article, we consider Nekbone, a mini-application for the CFD solver Nek5000, as a case study, and propose a methodology for enabling mixed-precision with the help of computer arithmetic tools and roofline model. We evaluate the derived mixed-precision program by combining metrics in three dimensions: accuracy, time-to-solution, and energy-to-solution. Notably, the introduction of mixed-precision in Nekbone, reducing time-to-solution by 40.7% and energy-to-solution by 47% on 128 MPI ranks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_11065 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enabling mixed-precision with the help of tools: A Nekbone case study Chen, Yanxiang Castro, Pablo de Oliveira Bientinesi, Paolo Iakymchuk, Roman Mathematical Software Distributed, Parallel, and Cluster Computing Software Engineering Mixed-precision computing has the potential to significantly reduce the cost of exascale computations, but determining when and how to implement it in programs can be challenging. In this article, we consider Nekbone, a mini-application for the CFD solver Nek5000, as a case study, and propose a methodology for enabling mixed-precision with the help of computer arithmetic tools and roofline model. We evaluate the derived mixed-precision program by combining metrics in three dimensions: accuracy, time-to-solution, and energy-to-solution. Notably, the introduction of mixed-precision in Nekbone, reducing time-to-solution by 40.7% and energy-to-solution by 47% on 128 MPI ranks. |
| title | Enabling mixed-precision with the help of tools: A Nekbone case study |
| topic | Mathematical Software Distributed, Parallel, and Cluster Computing Software Engineering |
| url | https://arxiv.org/abs/2405.11065 |