• Home
  • Search
  • Introducing Control Flow into Vectorized Code
  • Cite Icon39
  • https://doi.org/10.1109/pact.2007.41Copy DOI Icon

Introducing Control Flow into Vectorized Code

  • Sep 15, 2007
  • Jaewook Shin
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Single instruction multiple data (SIMD) functional units are ubiquitous in modern microprocessors. Effective use of these SIMD functional units is essential in achieving the highest possible performance. Automatic generation of SIMD instructions in the presence of control flow is challenging, however, not only because SIMD code is hard to generate in the presence of arbitrarily complex control flow, but also because the SIMD code executing the instructions in all control paths may slow compared to the scalar original, which may bypass a large portion of the code. One promising technique introduced recently involves inserting branches-on-superword-condition-codes (BOSCCs) to bypass vector instructions. In this paper, we describe two techniques that improve on the previous approach. First, BOSCCs are generated in a nested fashion so that even BOSCCs themselves can be bypassed by other BOSCCs. Second, we generate all vec_any_* instructions to bypass even some predicate-defining instructions. We implemented these techniques in a vectorizing compiler. On 14 kernels, the compiler achieves distinct speedups, including 1.99X over the previous technique that generates single- level BOSCCs and vec_any_ne only.

Similar Papers
  • Research Article
  • Citations1

Single instruction multiple data code auto generation for a very long instruction words digital signal processor in sensor‐based systems

  • Jun 01, 2013
  • IET Wireless Sensor Systems
  • Xu Yang +4
  • Conference Article
  • Citations6

Customized SIMD unit synthesis for system on programmable chip

  • Jan 01, 2006
  • Muhammad Omer Cheema +1
  • Conference Article
  • Citations3

Combining Run-Time Checks and Compile-Time Analysis to Improve Control Flow Auto-Vectorization

  • Oct 08, 2022
  • Bangtian Liu +4
  • Conference Article
  • Citations104

From SODA to scotch: The evolution of a wireless baseband processor

  • Nov 01, 2008
  • Mark Woh +10
  • Conference Article
  • Citations38

Efficient SIMD code generation for irregular kernels

  • Feb 25, 2012
  • Seonggun Kim +1
  • Research Article
  • Citations17

Efficient SIMD code generation for irregular kernels

  • Feb 25, 2012
  • ACM SIGPLAN Notices
  • Seonggun Kim +1
  • Book Chapter

Performance Analysis of Existing SIMD Architectures

  • Jan 01, 2019
  • Chao Cui +2
  • Conference Article
  • Citations95

Divergence Analysis and Optimizations

  • Oct 01, 2011
  • Bruno Coutinho +3
  • Research Article
  • Citations3

FastNBL: fast neighbor lists establishment for molecular dynamics simulation based on bitwise operations

  • Apr 29, 2019
  • The Journal of Supercomputing
  • Kun Li +4
  • Conference Article
  • Citations49

Using SIMD instructions for fast likelihood calculation in LVCSR

  • Jun 05, 2000
  • S Kanthak +2
  • Research Article
  • Citations4

Exploiting SIMD Instructions in Modern Microprocessors to Optimize the Performance of Stream Ciphers

  • May 16, 2013
  • International Journal of Computer Network and Information Security
  • P Mabin Joseph +3
  • Conference Article
  • Citations3

On the scalability of SIMD processing for software defined radio algorithms

  • Jul 01, 2010
  • Peter Westermann +1
  • Conference Article
  • Citations5

Design of a custom vector operation API exploiting SIMD intrinsics within Java

  • May 01, 2010
  • Jonathan Parri +4
  • Conference Article

Accelearation of Full-Search Algorithm on SIMD Architectures by Using Eight-Bit Partial Sums of Four Luminance Values

  • Dec 01, 2006
  • C J Duanmu
  • Conference Article
  • Citations2

Accelerating higher-order masking of AES using composite field and SIMD

  • Dec 01, 2015
  • Abdulaziz Miyajan +3
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.