Reproducible Triangular Solvers for High-Performance Computing

Roman Iakymchuk 1, 2 David Defour 3 Sylvain Collange 4 Stef Graillat 1
1 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 : On modern parallel architectures, floating-point computations may become non-deterministic and, therefore, non-reproducible mainly due to non-associativity of floating-point operations. We propose an algorithm to solve dense triangular systems by leveraging the standard parallel triangular solver and our, recently introduced, multi-level exact summation approach. Finally, we present implementations of the proposed fast reproducible triangular solver and results on recent NVIDIA GPUs.
Document type :
Conference papers
Complete list of metadatas

https://hal-lirmm.ccsd.cnrs.fr/lirmm-01206371
Contributor : David Defour <>
Submitted on : Monday, September 28, 2015 - 9:34:54 PM
Last modification on : Thursday, March 21, 2019 - 1:19:50 PM

Links full text

Identifiers

Citation

Roman Iakymchuk, David Defour, Sylvain Collange, Stef Graillat. Reproducible Triangular Solvers for High-Performance Computing. ITNG: Information Technology - New Generations, Apr 2015, Las Vegas, NV, United States. pp.353-358, ⟨10.1109/ITNG.2015.63⟩. ⟨lirmm-01206371⟩

Share

Metrics

Record views

472