1
�-�- EM Programmer's Notebook John L. Volakis ElectroScience Lab EJecrical Engineering Dept. The Ohio State University 1320 Kinnear Rd. Columbus, OH 43212 +1 (614) 292-5846 Tel. +1 (614) 292-7297 (Fax) volakis.1@osu.edu (e-mail) . David B. Davidson Dept. E&E Engineering Universi of Stellenbosch Stellenbosch 7600, South Africa (+27) 21 808 4458 (+27) 21 808 4981 (Fax) davidson@ing.sun.ac.za (e-mail) Foreword by the Editors Fiſteen years back, one us of (DBD) was performing doctoral work on using what was then a pioneering high-speed processing element, the transputer, for computational electromagnetics simu- lations. The aim was to exploit the power of the purpose-built tran- sputer hardware for high-speed computation - high-speed when compared to then-contemporary general-puose CPUs, such as the Intel 286s, 386s, etc., that is. (More details on the transputer, as well as some of its applications in early parallel processing sys- tems, may be found in [1, p. 211-212]). However, the phenomenal increase in CPU clock speeds throughout the 1990s killed off most competing technologies, including the ansputer. At present, the rate of increase of general CPU capability shows signs of flattening off somewhat, for a variety of techno- logical reasons. Interestingly, the niche in "computer ecology" that the transputer filled is now reappearing in response to this trend, but now using either proble gate aays, or aphical proc- essing units (GPUs). The latter offer exordinary floating-point speed - albeit for very specialized purposes - d this month's column discusses how GPUs can be used for one of the most read- 70 ily adapted algorithms in CEM, t he FDTD.We thank the authors for their interesting conibution. On a different topic, we would like to draw the reader's attention to the 8th Inteational Workshop on Finite Elements for Microwave Engineering (http://courses.ee.sun.ac.zEM20061); a call for papers appeared in the previous three issues of the Maga- zine. (Readers may recall a report on the sixth workshop in Chios, Greece, in the June, 2002, column). For the first time, the work- shop will be taking place in the southe hemisphere: in Stellen- bosch, South Aica, May 25-26, 2006. Reference 1. D. B. Davidson, Computational Electromagnetics f or and Microwave Engineering, Cambridge, UK, Cbridge Universi Press, 2005. IEEE Antennas and Ppagation Magazine, Vol. 47, No.6, December 2005

EM Programmer's Notebook

  • Upload
    amelia

  • View
    225

  • Download
    0

Embed Size (px)

Citation preview

Page 1: EM Programmer's Notebook

��

�-�-EM Programmer's Notebook

John L. Volakis ElectroScience Lab EJecrical Engineering Dept. The Ohio State University 1320 Kinnear Rd. Columbus, OH 43212 +1 (614) 292-5846 Tel. +1 (614) 292-7297 (Fax) volakis.1 @osu.edu (e-mail)

�,."..

David B. Davidson Dept. E&E Engineering University of Stellenbosch Stellenbosch 7600, South Africa (+27) 21 808 4458 (+27) 21 808 4981 (Fax) [email protected] (e-mail)

Foreword by the Editors

Fifteen years back, one us of (DBD) was performing doctoral work on using what was then a pioneering high-speed processing element, the transputer, for computational electromagnetics simu­lations. The aim was to exploit the power of the purpose-built tran­sputer hardware for high-speed computation - high-speed when compared to then-contemporary general-purpose CPUs, such as the Intel 286s, 386s, etc., that is. (More details on the transputer, as well as some of its applications in early parallel processing sys­tems, may be found in [1, p. 211-212]). However, the phenomenal increase in CPU clock speeds throughout the 1990s killed off most competing technologies, including the transputer.

At present, the rate of increase of general CPU capability shows signs of flattening off somewhat, for a variety of techno­logical reasons. Interestingly, the niche in "computer ecology" that the transputer filled is now reappearing in response to this trend, but now using either programmable gate arrays, or graphical proc­essing units (GPUs). The latter offer extraordinary floating-point speed - albeit for very specialized purposes - and this month's

column discusses how GPUs can be used for one of the most read-

70

ily adapted algorithms in CEM, the FDTD.We thank the authors for their interesting contribution.

On a different topic, we would like to draw the reader's attention to the 8th International Workshop on Finite Elements for Microwave Engineering (http://courses.ee.sun.ac.zalFEM20061); a call for papers appeared in the previous three issues of the Maga­zine. (Readers may recall a report on the sixth workshop in Chios, Greece, in the June, 2002, column). For the first time, the work­shop will be taking place in the southern hemisphere: in Stellen­bosch, South Africa, May 25-26, 2006.

Reference

1. D. B. Davidson, Computational Electromagnetics for RF and Microwave Engineering, Cambridge, UK, Cambridge University Press, 2005.

IEEE Antennas and Propagation Magazine, Vol. 47, No.6, December 2005