Improving workload balance and code optimization on processor-in-memory systems

被引：1

作者：

Chu, SL ^{[1
]}

Huang, TC ^{[1
]}

Lee, LC ^{[1
]}

机构：

[1] Natl Sun Yat Sen Univ, Dept Elect Engn, Kaohsiung 804, Taiwan

来源：

JOURNAL OF SYSTEMS AND SOFTWARE | 2004年 / 71卷 / 1-2期

关键词：

D O I：

10.1016/S0164-1212(02)00150-4

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

Processor-in-memory (PIM) architectures have recently been proposed, with the objective of reducing the performance gap between processor and memory. An earlier study of Huang and Chu [Proceedings of 2nd Workshop on Intelligent Memory Systems, Cambridge, MA, 2000] designed a statement-based parallelizing system, SAGE, to exploit the potential benefits of PIM. This study extends this system to achieve better performance. Several comprehensive optimization approaches, including self-patch weight evaluation, loop splitting for PIM, intelligent memory operation (IMOP) recognition, and tiling for PIM, are devised to produce execution schedules with improved load balance. Experimental results confirm the effectiveness of the proposed method. (C) 2002 Elsevier Inc. All rights reserved.

引用

页码：71 / 82

页数：12

共 14 条

[1] Combining optimization for cache and instruction-level parallelism [J].

Carr, S .

PROCEEDINGS OF THE 1996 CONFERENCE ON PARALLEL ARCHITECTURES AND COMPILATION TECHNIQUES (PACT '96), 1996, :238-247

[2]

GRANACKI J, 1998, DATA INTENSIVE ARCHI

[3]

HUANG TC, 2001, P ISCA 16 INT C COMP, P452

[4]

HUANG TC, 2000, P 2 WORKSH INT MEM S, P12

[5]

JIMENEZ M, 1999, THESIS U POLITECNICA

[6]

JUDD D, 2000, P 2 WORKSH INT MEM S

[7]

KANG Y, 1999, P INT C COMP DES ICD

[8]

MORITZ CA, 2000, P 2 WORKSH INT MEM S

[9] Active pages: A computation model for intelligent memory [J].

Oskin, M ;

Chong, FT ;

Sherwood, T .

25TH ANNUAL INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE, PROCEEDINGS, 1998, :192-203

[10]

PATTERSON D, 1997, CASE INTELLIGENT DRA, P33

← 1 2 →