Implementation and Efficiency of Reproducible Level 1 BLAS

Chemseddine Chohra 1 Philippe Langlois 1 David Parello 1
1 DALI - Digits, Architectures et Logiciels Informatiques
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier, UPVD - Université de Perpignan Via Domitia
Abstract : Numerical reproducibility failures appear in massively parallel floating-point computations. One way to guarantee this reproducibility is to extend the IEEE-754 correct rounding to larger computing sequences, e.g. to the BLAS. Is the extra cost for numerical reproducibility acceptable in practice? We present solutions and experiments for the level 1 BLAS. We detail optimized implementations and we conclude about their efficiency.
Complete list of metadatas

Cited literature [13 references]  Display  Hide  Download

https://hal-lirmm.ccsd.cnrs.fr/lirmm-01179986
Contributor : Philippe Langlois <>
Submitted on : Thursday, July 23, 2015 - 4:43:08 PM
Last modification on : Tuesday, February 19, 2019 - 8:28:01 PM
Long-term archiving on : Saturday, October 24, 2015 - 11:44:20 AM

File

15 pages.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : lirmm-01179986, version 1

Collections

Citation

Chemseddine Chohra, Philippe Langlois, David Parello. Implementation and Efficiency of Reproducible Level 1 BLAS. [Research Report] DALI - UPVD/LIRMM, UCD. 2015. ⟨lirmm-01179986⟩

Share

Metrics

Record views

308

Files downloads

951