Record Details

FPGA based high performance double-precision matrix multiplication

DSpace at IIT Bombay

View Archive Info
 
 
Field Value
 
Title FPGA based high performance double-precision matrix multiplication
 
Creator KUMAR, VBY
JOSHI, S
PATKAR, SB
NARAYANAN, H
 
Subject high performance computing
matrix multiplication
rank-1 scheme
fpga implementation
memory-bandwidth trade-off
scalability
 
Description We present two designs (I and II) for IEEE 754 double precision floating point matrix multiplication, optimized for implementation on high-end FPGAs. It forms the kernel in many important tile-based BLAS algorithms, making an excellent candidate for acceleration. The designs, both based on the rank-1 update scheme, can handle arbitrary matrix sizes, and are able to sustain their peak performance except during an initial latency period. Through these designs, the trade-offs involved in terms of local-memory and bandwidth for an FPGA implementation are demonstrated and an analysis is presented for the optimal choice of design parameters. The designs, implemented on a Virtex-5 SX240T FPGA, scale gracefully from 1 to 40 processing elements(PEs) with a less than 1% degradation in the design frequency of 373 MHz. With 40 PEs and a design speed of 373 MHz, a sustained performance of 29.8 GFLOPS is possible with a bandwidth requirement of 750 MB/s for design-II and 5.9 GB/s for design-I. This compares favourably with both related art and general purpose CPU implementations.
 
Publisher SPRINGER/PLENUM PUBLISHERS
 
Date 2011-10-22T07:22:36Z
2011-12-15T09:10:48Z
2011-10-22T07:22:36Z
2011-12-15T09:10:48Z
2010
 
Type Article; Proceedings Paper
 
Identifier INTERNATIONAL JOURNAL OF PARALLEL PROGRAMMING,38,322-338
0885-7458
http://dx.doi.org/10.1007/s10766-010-0131-8
http://dspace.library.iitb.ac.in/xmlui/handle/10054/14840
http://hdl.handle.net/100/1667
 
Source 22nd International Conference on VLSI Design/8th International Conference on Embedded Systems,New Delhi, INDIA,JAN, 2009
 
Language English