Non-Preconditioned Conjugate Gradient on Cell and FPGA based Hybrid Supercomputer Nodes

David DuBois; Andrew DuBois; Thomas Boorman; Carolyn Connor; IEEE Computer Soc

doi:10.1109/FCCM.2009.26

Back

Conference proceeding

Non-Preconditioned Conjugate Gradient on Cell and FPGA based Hybrid Supercomputer Nodes

David DuBois, Andrew DuBois, Thomas Boorman, Carolyn Connor and IEEE Computer Soc

PROCEEDINGS OF THE 2009 17TH IEEE SYMPOSIUM ON FIELD PROGRAMMABLE CUSTOM COMPUTING MACHINES, pp.201-208

Annual IEEE Symposium on Field-Programmable Custom Computing Machines

01/01/2009

DOI: https://doi.org/10.1109/FCCM.2009.26

Abstract and subjects

Computer Science

Computer Science, Hardware & Architecture

Science & Technology

Technology

This work presents a detailed implementation of a double precision, non-preconditioned, Conjugate Gradient algorithm on a Roadrunner heterogeneous supercomputer node. These nodes utilize the Cell Broadband Engine Architecture (TM) in conjunction with x86 Opteron (TM) processors from AMD. We implement a common Conjugate Gradient algorithm, on a variety of systems, to compare and contrast performance. Implementation results are presented for the Roadrunner hybrid supercomputer, SRC Computers, Inc. MAPStation SRC-6 FPGA enhanced hybrid supercomputer, and AMD Opteron only. In all hybrid implementations wall clock time is measured, including all transfer overhead and compute timings.

Metrics

3 Record Views

Details

Title: Non-Preconditioned Conjugate Gradient on Cell and FPGA based Hybrid Supercomputer Nodes
Authors/Creators - name: David DuBois
Andrew DuBois
Thomas Boorman
Carolyn Connor
IEEE Computer Soc
Publication Details: PROCEEDINGS OF THE 2009 17TH IEEE SYMPOSIUM ON FIELD PROGRAMMABLE CUSTOM COMPUTING MACHINES, pp.201-208
Series: Annual IEEE Symposium on Field-Programmable Custom Computing Machines
Publisher: IEEE
Number of pages: 8
Language: English
Resource Type: Conference proceeding
ISBN: 9780769537160; 0769537162
DOI: https://doi.org/10.1109/FCCM.2009.26