SPIRAL:: Code generation for DSP transforms

Markus Püschel

doi:10.1109/JPROC.2004.840306

2005-02-01

505

PRE

DOI：10.1109/JPROC.2004.840306

原文链接

摘要

En 中文

Fast changing, increasingly complex, and diverse computing platforms pose central problems in scientific computing: How to achieve, with reasonable effort, portable optimal performance? We present SPIRAL, which considers this problem for the performance-critical domain of linear digital signal processing (DSP) transforms. For a specified transform, SPIRAL automatically gen-crates high-performance code that is tuned to the given platform. SPIRAL formulates the timing as an optimization problem and exploits the domain-specific mathematical structure of transform algorithins to implement a feedback-driven optimizer Similar to a human expert, for a specified transform, SPIRAL intelligently generates and explores algorithmic and implementation choices to find the best match to the computer's microarchitecture. The intelligence is provided by search and learning techniques that exploit the structure of the algorithm and implementation space to guide the exploration and optimization. SPIRAL generates high-performance code for a broad set of DSP transforms, including the discrete Fourier transform, other trigonometric transforms, filter transforms, and discrete wavelet transforms. Experimental results show that the code generated by SPIRAL competes with, and sometimes outperforms, the best available human timed transform library code.

Keyword:

adaptation

automatic peformance tuning

code optimization

discrete cosine transform (DCT)

discrete Fourier transform (DFT)

fast Fourier transform (FFT)

filter

genetic and evolutionary algorithm

high-performance computing

learning

library generation

linear signal transform

Markov decision process

wavelet

AI总结

对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Proceedings of the IEEE

IF:

25.9

论文数:

10.0K

被引数:

4.5W

机构

暂无机构信息

引用论文

Parallel VSIPL++: An open standard software library for high-performance parallel signal processing

PROCEEDINGS OF THE IEEE

IF25.9

2005-02-01

PREAI

Lebak, J; Kepner, J; Hoffmann, H; Rutledge, E

Synthesis of high-performance parallel programs for a class of Ab Initio quantum chemistry models 一类从头算量子化学模型的高性能并行程序的合成

PROCEEDINGS OF THE IEEE

IF25.9

2005-02-01

139

PREAI

Baumgartner, G; Auer, AA; Bernholdt, DE; Bibireata, A; Choppella, V; Cociorva, D; Gao, XY; Harrison, RJ; Hirata, S; Krishnamoorthy, S; Krishnan, S; Lam, CC; Lu, QD; Nooijen, M; Pitzer, RM; Ramanujam, J; Sadayappan, P; Sibiryakov, A

Highly Selective Hydroboration of Terminal Alkenes Catalyzed by a Cobalt Pincer Complex Featuring a Central Reactive N-Heterocyclic Phosphido Fragment

Organometallics

IF0

2021-04-07

PREAI

Andrew M. Poitras; Leah K. Oliemuller; Gillian P. Hatzis; Christine M. Thomas

Vioxx: an unequal partnership between safety and efficacy

The Lancet

IF0

2004-10-01