Avoiding data conversions in embedded media processors

Authors:
Ben Juurlink;Asadollah Shahbahrami;Stamatis Vassiliadis
Affiliations:
Delft University of Technology, The Netherlands;Delft University of Technology, The Netherlands;Delft University of Technology, The Netherlands
Venue:
Proceedings of the 2005 ACM symposium on Applied computing
Year:
2005

Citing 2
Cited 3

Reconfigurable media processing

Parallel Computing - Parallel computing in image and video processing
Measuring the Performance of Multimedia Instruction Sets

IEEE Transactions on Computers

Matrix register file and extended subwords: two techniques for embedded media processors

Proceedings of the 2nd conference on Computing frontiers
Avoiding conversion and rearrangement overhead in SIMD architectures

International Journal of Parallel Programming
Limitations of special-purpose instructions for similarity measurements in media SIMD extensions

CASES '06 Proceedings of the 2006 international conference on Compilers, architecture and synthesis for embedded systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Complex application-specific media instructions and kernels are emulated with simple to implement extended subword instructions. We show that assuming extended register file entries to accommodate intermediate results and by implementing a few simple instructions, packing/unpacking, saturation, and frequently used complex instructions can be practically eliminated. It is shown that in most emulations there is a potential performance improvement, making the proposed scheme suitable for embedded processors with a limited hardware budget.