Abstract
Computing many small 2D convolutions using FFTs is a basis for a large number of applications in many domains in science and engineering, among them electromagnetic diffraction modeling in physics. The GPU architecture seems to be a suitable architecture to accelerate these convolutions, but reaching high application performance requires substantial development time and non-portable optimizations. In this work, we present the techniques, performance results and considerations to accelerate small 2D convolutions using CUDA, and compare performance to a multi-threaded CPU implementation. To improve programmability and performance of applications that make heavy use of small convolutions, we argue that two improvements to software and hardware are needed: FFT libraries must be extended with a single convolution function and communication bandwidth between CPU and GPU needs to be drastically improved.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Govindaraju, N., Lloyd, B., Dotsenko, Y., Smith, B., Manferdelli, J.: High performance discrete fourier transforms on graphics processors. In: Proc. of the ACM/IEEE Conf. on Supercomputing, pp. 1–12. IEEE Press, Los Alamitos (2008)
Podlozhnyuk, V.: Image convolution with CUDA. Tech. rep., NVIDIA (2007)
NVIDIA: CUDA Programming Guide (February 2010)
Podlozhnyuk, V.: FFT-based 2D convolution. Tech. rep., NVIDIA (2007)
Podlozhnyuk, V.: Image convolution with CUDA. Tech. rep., NVIDIA (2007)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Al Umairy, S.A.H., van Amesfoort, A.S., Setija, I.D., van Beurden, M.C., Sips, H.J. (2011). On the Use of Small 2D Convolutions on GPUs. In: Varbanescu, A.L., Molnos, A., van Nieuwpoort, R. (eds) Computer Architecture. ISCA 2010. Lecture Notes in Computer Science, vol 6161. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-24322-6_6
Download citation
DOI: https://doi.org/10.1007/978-3-642-24322-6_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-24321-9
Online ISBN: 978-3-642-24322-6
eBook Packages: Computer ScienceComputer Science (R0)