A GPU Spatial Processing System for CHIME

We present an overview of the Graphics Processing Unit (GPU) based spatial processing system created for the Canadian Hydrogen Intensity Mapping Experiment (CHIME). The design employs AMD S9300x2 GPUs and readily-available commercial hardware in its processing nodes to provide a cost- and power-efficient processing substrate. These nodes are supported by a liquid-cooling system which allows continuous operation with modest power consumption and in all but the most adverse conditions. Capable of continuously correlating 2048 receiver-polarizations across 400\,MHz of bandwidth, the CHIME X-engine constitutes the most powerful radio correlator currently in existence. It receives $6.6$\,Tb/s of channelized data from CHIME's FPGA-based F-engine, and the primary correlation task requires $8.39\times10^{14}$ complex multiply-and-accumulate operations per second. The same system also provides formed-beam data products to commensal FRB and Pulsar experiments; it constitutes a general spatial-processing system of unprecedented scale and capability, with correspondingly great challenges in computation, data transport, heat dissipation, and interference shielding.

[1]  C. Carilli,et al.  Synthesis Imaging in Radio Astronomy II , 1999 .

[2]  Michael A. Clark,et al.  Accelerating radio astronomy cross-correlation with graphics processing units , 2011, Int. J. High Perform. Comput. Appl..

[3]  Robert J. Selina,et al.  The Next-Generation Very Large Array: a technical overview , 2018, Astronomical Telescopes + Instrumentation.

[4]  Meiling Deng,et al.  The cloverleaf antenna: A compact wide-bandwidth dual-polarization feed for CHIME , 2014, 2014 16th International Symposium on Antenna Technology and Applied Electromagnetics (ANTEM).

[5]  Matias Zaldarriaga,et al.  Fast Fourier transform telescope , 2008, 0805.4414.

[6]  G. Swenson,et al.  Interferometry and Synthesis in Radio Astronomy , 1986 .

[7]  Nolan Thomas Denman Digital Signal Processing for the Canadian Hydrogen Intensity Mapping Experiment , 2019 .

[8]  Matias Zaldarriaga,et al.  Omniscopes: Large area telescope arrays with only NlogN computational cost , 2009, 0909.0001.

[9]  Graeme Smecher,et al.  Canadian Hydrogen Intensity Mapping Experiment (CHIME) pathfinder , 2014, Astronomical Telescopes and Instrumentation.

[10]  Thomas A. Limoncelli,et al.  The Practice of System and Network Administration , 2001 .

[11]  B. R. Barsdell,et al.  Digital Signal Processing using Stream High Performance Computing: A 512-input Broadband Correlator for Radio Astronomy , 2014, 1411.3751.

[12]  Abraham Loeb,et al.  Possibility of precise measurement of the cosmological power spectrum with a dedicated survey of 21 cm emission after reionization. , 2008, Physical review letters.

[13]  Stephanie Thalberg,et al.  Interferometry And Synthesis In Radio Astronomy , 2016 .

[14]  Peter Klages,et al.  GPU kernels for high-speed 4-bit astrophysical data processing , 2015, 2015 IEEE 26th International Conference on Application-specific Systems, Architectures and Processors (ASAP).