Solving the Table Maker's Dilemma on Current SIMD Architectures - Sorbonne Université Access content directly
Journal Articles Scalable Computing : Practice and Experience Year : 2016

Solving the Table Maker's Dilemma on Current SIMD Architectures

Abstract

Correctly-rounded implementations of some elementary functions are recommended by the IEEE 754-2008 standard, which aims at ensuring portable and predictable floating-point computations. Such implementations require the solving of the Table Maker's Dilemma which implies a huge amount of computation time. These computations are embarrassingly and massively parallel, but present control flow divergence which limits performance at the SIMD parallelism level, whose share in the overall performance of current and forthcoming HPC architectures is increasing. In this paper, we show that efficiently solving the Table Maker's Dilemma on various multi-core and many-core SIMD architectures (CPUs, GPUs, Intel Xeon Phi) requires to jointly handle divergence at the algorithmic, programming and hardware levels in order to scale with the number of SIMD lanes. Depending on the architecture, the performance gains can reach 10.5x over divergent code, or be constrained by different limits that we detail.
Fichier principal
Vignette du fichier
Solving.pdf (400.94 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-01356850 , version 1 (17-03-2024)

Identifiers

Cite

Christophe Avenel, Pierre Fortin, Mourad Gouicem, Zaidi Samia. Solving the Table Maker's Dilemma on Current SIMD Architectures. Scalable Computing : Practice and Experience, 2016, 17 (3), pp.237-249. ⟨10.12694/scpe.v17i3.1183⟩. ⟨hal-01356850⟩
470 View
8 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More