Enabling a High Throughput Real Time Data Pipeline for a Large Radio Telescope Array with GPUs
Initiative in Innovative Computing, 29 Oxford Street, Cambridge MA 02138
Computer Physics Communications, Volume 181, Issue 10, p. 1707-1714, arXiv:1003.5575 [astro-ph.IM] (29 Mar 2010)
@article{edgar2010enabling,
title={Enabling a high throughput real time data pipeline for a large radio telescope array with GPUs},
author={Edgar, RG and Clark, MA and Dale, K. and Mitchell, DA and Ord, SM and Wayth, RB and Pfister, H. and Greenhill, LJ},
journal={Computer Physics Communications},
issn={0010-4655},
year={2010},
publisher={Elsevier}
}
The Murchison Widefield Array (MWA) is a next-generation radio telescope currently under construction in the remote Western Australia Outback. Raw data will be generated continuously at 5GiB/s, grouped into 8s cadences. This high throughput motivates the development of on-site, real time processing and reduction in preference to archiving, transport and off-line processing. Each batch of 8s data must be completely reduced before the next batch arrives. Maintaining real time operation will require a sustained performance of around 2.5TFLOP/s (including convolutions, FFTs, interpolations and matrix multiplications). We describe a scalable heterogeneous computing pipeline implementation, exploiting both the high computing density and FLOP-per-Watt ratio of modern GPUs. The architecture is highly parallel within and across nodes, with all major processing elements performed by GPUs. Necessary scatter-gather operations along the pipeline are loosely synchronized between the nodes hosting the GPUs. The MWA will be a frontier scientific instrument and a pathfinder for planned peta- and exascale facilities.
November 12, 2010 by hgpu
