Intra-vector SIMD instructions for core specialization

Authors:
Cor Meenderinck;Ben Juurlink
Affiliations:
Computer Engineering Laboratory, Faculty of Electrical Engineering, Mathematics, and Computer Science, Delft University of Technology, Delft, The Netherlands;Computer Engineering Laboratory, Faculty of Electrical Engineering, Mathematics, and Computer Science, Delft University of Technology, Delft, The Netherlands
Venue:
ICCD'09 Proceedings of the 2009 IEEE international conference on Computer design
Year:
2009

Citing 4
Cited 0

Low power DSP's for wireless communications (embedded tutorial session)

ISLPED '00 Proceedings of the 2000 international symposium on Low power electronics and design
SODA: A High-Performance DSP Architecture for Software-Defined Radio

IEEE Micro
Optimizing JPEG2000 Still Image Encoding on the Cell Broadband Engine

ICPP '08 Proceedings of the 2008 37th International Conference on Parallel Processing
Specialization of the Cell SPE for Media Applications

ASAP '09 Proceedings of the 2009 20th IEEE International Conference on Application-specific Systems, Architectures and Processors

Quantified Score

Hi-index	0.00

Visualization

Abstract

Current research is mainly focussing on exploiting TLP to increase performance. Another avenue, however, for achieving performance scalability is specialization. In this paper we propose application specific intra-vector instructions for two dimensional signal processing kernels. In such kernels usually significant data rearrangement overhead is required in order to use the SIMD capabilities. When using the intra-vector instructions the overhead can be avoided. We have implemented intra-vector instructions in the Cell SPU core and measured speedups of up to 2.06, with an average of 1.45.