Near Memory Acceleration on High Resolution Radio Astronomy Imaging

Stefano Corda; Bram Veenboer; Ahsan Javed Awan; Akash Kumar; Roel Jordans; Henk Corporaal

doi:10.1109/MECO49872.2020.9134089

Near Memory Acceleration on High Resolution Radio Astronomy Imaging

Stefano Corda, Bram Veenboer, Ahsan Javed Awan, Akash Kumar, Roel Jordans, Henk Corporaal

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

5 Citations (Scopus)

Abstract

Modern radio telescopes like the Square Kilometer Array (SKA) will need to process in real-time exabytes of radio-astronomical signals to construct a high-resolution map of the sky. Near-Memory Computing (NMC) could alleviate the performance bottlenecks due to frequent memory accesses in a state-of-the-art radio-astronomy imaging algorithm. In this paper, we show that a sub-module performing a two-dimensional fast Fourier transform (2D FFT) is memory bound using CPI breakdown analysis on IBM Power9. Then, we present an NMC approach on FPGA for 2D FFT that outperforms a CPU by up to a factor of 120x and performs comparably to a high-end GPU, while using less bandwidth and memory.

Original language	English
Title of host publication	2020 9th Mediterranean Conference on Embedded Computing (MECO)
Publisher	Institute of Electrical and Electronics Engineers
Number of pages	6
ISBN (Electronic)	978-1-7281-6949-1
DOIs	https://doi.org/10.1109/MECO49872.2020.9134089
Publication status	Published - 7 Jul 2020
Event	9th Mediterranean Conference on Embedded Computing, MECO 2020 - Budva, Montenegro Duration: 8 Jun 2020 → 11 Jun 2020

Conference

Conference	9th Mediterranean Conference on Embedded Computing, MECO 2020
Country/Territory	Montenegro
City	Budva
Period	8/06/20 → 11/06/20

Funding

This work is funded by the European Commission under Marie Sklodowska-Curie Innovative Training Networks European Industrial Doctorate (Project ID: 676240). We would like to thank Jan van Lunteren from IBM Research for providing the Access Processor and NMC accelerator architecture, and Sambit Nayak from Ericsson Research for his feedback on the draft of the paper.

Funders	Funder number
European Union's Horizon 2020 - Research and Innovation Framework Programme	676240
European Commission

Access to Document

10.1109/MECO49872.2020.9134089

Cite this

@inproceedings{91127228192a45e7bc2bb3c5cfa888fd,

title = "Near Memory Acceleration on High Resolution Radio Astronomy Imaging",

abstract = "Modern radio telescopes like the Square Kilometer Array (SKA) will need to process in real-time exabytes of radio-astronomical signals to construct a high-resolution map of the sky. Near-Memory Computing (NMC) could alleviate the performance bottlenecks due to frequent memory accesses in a state-of-the-art radio-astronomy imaging algorithm. In this paper, we show that a sub-module performing a two-dimensional fast Fourier transform (2D FFT) is memory bound using CPI breakdown analysis on IBM Power9. Then, we present an NMC approach on FPGA for 2D FFT that outperforms a CPU by up to a factor of 120x and performs comparably to a high-end GPU, while using less bandwidth and memory.",

author = "Stefano Corda and Bram Veenboer and Awan, {Ahsan Javed} and Akash Kumar and Roel Jordans and Henk Corporaal",

year = "2020",

month = jul,

day = "7",

doi = "10.1109/MECO49872.2020.9134089",

language = "English",

booktitle = "2020 9th Mediterranean Conference on Embedded Computing (MECO)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "9th Mediterranean Conference on Embedded Computing, MECO 2020 ; Conference date: 08-06-2020 Through 11-06-2020",

}

Corda, S, Veenboer, B, Awan, AJ, Kumar, A, Jordans, R & Corporaal, H 2020, Near Memory Acceleration on High Resolution Radio Astronomy Imaging. in 2020 9th Mediterranean Conference on Embedded Computing (MECO)., 9134089, Institute of Electrical and Electronics Engineers, 9th Mediterranean Conference on Embedded Computing, MECO 2020, Budva, Montenegro, 8/06/20. https://doi.org/10.1109/MECO49872.2020.9134089

Near Memory Acceleration on High Resolution Radio Astronomy Imaging. / Corda, Stefano; Veenboer, Bram; Awan, Ahsan Javed et al.
2020 9th Mediterranean Conference on Embedded Computing (MECO). Institute of Electrical and Electronics Engineers, 2020. 9134089.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Near Memory Acceleration on High Resolution Radio Astronomy Imaging

AU - Corda, Stefano

AU - Veenboer, Bram

AU - Awan, Ahsan Javed

AU - Kumar, Akash

AU - Jordans, Roel

AU - Corporaal, Henk

PY - 2020/7/7

Y1 - 2020/7/7

N2 - Modern radio telescopes like the Square Kilometer Array (SKA) will need to process in real-time exabytes of radio-astronomical signals to construct a high-resolution map of the sky. Near-Memory Computing (NMC) could alleviate the performance bottlenecks due to frequent memory accesses in a state-of-the-art radio-astronomy imaging algorithm. In this paper, we show that a sub-module performing a two-dimensional fast Fourier transform (2D FFT) is memory bound using CPI breakdown analysis on IBM Power9. Then, we present an NMC approach on FPGA for 2D FFT that outperforms a CPU by up to a factor of 120x and performs comparably to a high-end GPU, while using less bandwidth and memory.

AB - Modern radio telescopes like the Square Kilometer Array (SKA) will need to process in real-time exabytes of radio-astronomical signals to construct a high-resolution map of the sky. Near-Memory Computing (NMC) could alleviate the performance bottlenecks due to frequent memory accesses in a state-of-the-art radio-astronomy imaging algorithm. In this paper, we show that a sub-module performing a two-dimensional fast Fourier transform (2D FFT) is memory bound using CPI breakdown analysis on IBM Power9. Then, we present an NMC approach on FPGA for 2D FFT that outperforms a CPU by up to a factor of 120x and performs comparably to a high-end GPU, while using less bandwidth and memory.

UR - http://www.scopus.com/inward/record.url?scp=85088539721&partnerID=8YFLogxK

U2 - 10.1109/MECO49872.2020.9134089

DO - 10.1109/MECO49872.2020.9134089

M3 - Conference contribution

AN - SCOPUS:85088539721

BT - 2020 9th Mediterranean Conference on Embedded Computing (MECO)

PB - Institute of Electrical and Electronics Engineers

T2 - 9th Mediterranean Conference on Embedded Computing, MECO 2020

Y2 - 8 June 2020 through 11 June 2020

ER -

Near Memory Acceleration on High Resolution Radio Astronomy Imaging

Abstract

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this