Reproducible and Accurate Matrix Multiplication

Roman Iakymchuk 1, 2 David Defour 3 Sylvain Collange 4 Stef Graillat 2
2 PEQUAN - Performance et Qualité des Algorithmes Numériques
LIP6 - Laboratoire d'Informatique de Paris 6
3 DALI - Digits, Architectures et Logiciels Informatiques
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier, UPVD - Université de Perpignan Via Domitia
4 ALF - Amdahl's Law is Forever
Inria Rennes – Bretagne Atlantique , IRISA-D3 - ARCHITECTURE
Abstract : Due to non-associativity of floating-point operations and dynamic scheduling on parallel architectures, getting a bit-wise reproducible floating-point result for multiple executions of the same code on different or even similar parallel architectures is challenging. In this paper, we address the problem of reproducibility in the context of matrix multiplication and propose an algorithm that yields both reproducible and accurate results. This algorithm is composed of two main stages: a filtering stage that uses fast vectorized floating-point expansions in conjunction with error-free transformations; an accumulation stage based on Kulisch long accumulators in a high-radix carry-save representation. Finally, we provide implementations and performance results in parallel environments like GPUs.
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01539180
Contributor : Stef Graillat <>
Submitted on : Wednesday, June 14, 2017 - 3:31:07 PM
Last modification on : Thursday, March 21, 2019 - 2:44:09 PM

Identifiers

Citation

Roman Iakymchuk, David Defour, Sylvain Collange, Stef Graillat. Reproducible and Accurate Matrix Multiplication. SCAN: Scientific Computing, Computer Arithmetic and Validated Numerics, Sep 2014, Wurzburg, Germany. pp.126-137, ⟨10.1007/978-3-319-31769-4_11⟩. ⟨hal-01539180⟩

Share

Metrics

Record views

533