Skip to Main content Skip to Navigation
Conference papers

Reproducible floating-point atomic addition in data-parallel environment

David Defour 1 Sylvain Collange 2
1 DALI - Digits, Architectures et Logiciels Informatiques
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier, UPVD - Université de Perpignan Via Domitia
2 ALF - Amdahl's Law is Forever
Inria Rennes – Bretagne Atlantique , IRISA-D3 - ARCHITECTURE
Abstract : Floating-point additions in concurrent execution environment are known to be hazardous, as the result depends on the order in which operations are performed. This problem is encountered in data parallel execution environments such as GPUs, where reproducibility involving floating-point atomic addition is challenging. This problem is due to the rounding error or cancellation that appears for each operation, combined with the lack of control over execution order. In this article we propose two solutions to address this problem: work reassignment and fixed-point accumulation. Work reassignment consists in enforcing an execution order that leads to weak reproducibility. Fixed-point accumulation consists in avoiding rounding error performed at each operation thanks to a long accumulator that leads to strong reproducibility.
Document type :
Conference papers
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01267755
Contributor : David Defour <>
Submitted on : Thursday, June 27, 2019 - 9:33:37 AM
Last modification on : Friday, July 10, 2020 - 4:19:44 PM

Links full text

Identifiers

Citation

David Defour, Sylvain Collange. Reproducible floating-point atomic addition in data-parallel environment. ACSIS, Sep 2015, Lodz, Poland. pp.721-728, ⟨10.15439/2015F86⟩. ⟨hal-01267755⟩

Share

Metrics

Record views

400