SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication

Qi Nie; Sharad Malik

doi:10.1109/ISCAS.2019.8702111

SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

To improve the performance of sparse matrix-matrix multiplication (SpMM) running on a specialized architecture, orchestrating a data flow that maximizes data reuse in local memory is critical but challenging due to the irregular non-zero element locations and the wide range of sparsity. In this work, we proposed SpFlow, a memory-driven data flow optimization framework for SpMM. SpFlow can realize 54X fewer DRAM accesses and 97X fewer SRAM accesses on average than a GPU running the cuSPARSE kernel. And in comparison with a state-of-the-art accelerator, the performance can be improved by 3X, and SRAM accesses reduced by 5X on average.

Original language	English (US)
Title of host publication	2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728103976
DOIs	https://doi.org/10.1109/ISCAS.2019.8702111
State	Published - Jan 1 2019
Event	2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Sapporo, Japan Duration: May 26 2019 → May 29 2019

Publication series

Name	Proceedings - IEEE International Symposium on Circuits and Systems
Volume	2019-May
ISSN (Print)	0271-4310

Conference

Conference	2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019
Country	Japan
City	Sapporo
Period	5/26/19 → 5/29/19

All Science Journal Classification (ASJC) codes

Electrical and Electronic Engineering

Access to Document

10.1109/ISCAS.2019.8702111

Fingerprint Dive into the research topics of 'SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication'. Together they form a unique fingerprint.

Cite this

Nie, Q., & Malik, S. (2019). SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication. In 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings [8702111] (Proceedings - IEEE International Symposium on Circuits and Systems; Vol. 2019-May). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISCAS.2019.8702111

@inproceedings{ef72e442bdcc4b6c9183d3b3a0983167,

title = "SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication",

abstract = "To improve the performance of sparse matrix-matrix multiplication (SpMM) running on a specialized architecture, orchestrating a data flow that maximizes data reuse in local memory is critical but challenging due to the irregular non-zero element locations and the wide range of sparsity. In this work, we proposed SpFlow, a memory-driven data flow optimization framework for SpMM. SpFlow can realize 54X fewer DRAM accesses and 97X fewer SRAM accesses on average than a GPU running the cuSPARSE kernel. And in comparison with a state-of-the-art accelerator, the performance can be improved by 3X, and SRAM accesses reduced by 5X on average.",

author = "Qi Nie and Sharad Malik",

year = "2019",

month = jan,

day = "1",

doi = "10.1109/ISCAS.2019.8702111",

language = "English (US)",

series = "Proceedings - IEEE International Symposium on Circuits and Systems",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings",

address = "United States",

note = "2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 ; Conference date: 26-05-2019 Through 29-05-2019",

}

Nie, Q & Malik, S 2019, SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication. in 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings., 8702111, Proceedings - IEEE International Symposium on Circuits and Systems, vol. 2019-May, Institute of Electrical and Electronics Engineers Inc., 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019, Sapporo, Japan, 5/26/19. https://doi.org/10.1109/ISCAS.2019.8702111

SpFlow : Memory-driven data flow optimization for sparse matrix-matrix multiplication. / Nie, Qi; Malik, Sharad.

2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. 8702111 (Proceedings - IEEE International Symposium on Circuits and Systems; Vol. 2019-May).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - SpFlow

T2 - 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019

AU - Nie, Qi

AU - Malik, Sharad

PY - 2019/1/1

Y1 - 2019/1/1

N2 - To improve the performance of sparse matrix-matrix multiplication (SpMM) running on a specialized architecture, orchestrating a data flow that maximizes data reuse in local memory is critical but challenging due to the irregular non-zero element locations and the wide range of sparsity. In this work, we proposed SpFlow, a memory-driven data flow optimization framework for SpMM. SpFlow can realize 54X fewer DRAM accesses and 97X fewer SRAM accesses on average than a GPU running the cuSPARSE kernel. And in comparison with a state-of-the-art accelerator, the performance can be improved by 3X, and SRAM accesses reduced by 5X on average.

AB - To improve the performance of sparse matrix-matrix multiplication (SpMM) running on a specialized architecture, orchestrating a data flow that maximizes data reuse in local memory is critical but challenging due to the irregular non-zero element locations and the wide range of sparsity. In this work, we proposed SpFlow, a memory-driven data flow optimization framework for SpMM. SpFlow can realize 54X fewer DRAM accesses and 97X fewer SRAM accesses on average than a GPU running the cuSPARSE kernel. And in comparison with a state-of-the-art accelerator, the performance can be improved by 3X, and SRAM accesses reduced by 5X on average.

UR - http://www.scopus.com/inward/record.url?scp=85066812084&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066812084&partnerID=8YFLogxK

U2 - 10.1109/ISCAS.2019.8702111

DO - 10.1109/ISCAS.2019.8702111

M3 - Conference contribution

AN - SCOPUS:85066812084

T3 - Proceedings - IEEE International Symposium on Circuits and Systems

BT - 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 26 May 2019 through 29 May 2019

ER -

Nie Q, Malik S. SpFlow: Memory-driven data flow optimization for sparse matrix-matrix multiplication. In 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. 8702111. (Proceedings - IEEE International Symposium on Circuits and Systems). https://doi.org/10.1109/ISCAS.2019.8702111