Alleviate Chip Pin Constraint for Multicore Processor by On/Off-Chip Power Delivery System Codesign

被引:0
作者
Wang, Xuan [1 ]
Xu, Jiang [1 ]
Wang, Zhe [1 ]
Li, Haoran [1 ]
Wang, Zhehui [1 ]
Yang, Peng [1 ]
Duong, Luan H. K. [1 ]
Maeda, Rafael K. V. [1 ]
Wang, Zhifei [1 ]
机构
[1] Hong Kong Univ Sci Technol, Hong Kong, Hong Kong, Peoples R China
关键词
Power delivery system; pin count constraint; on-chip voltage regulator; analytical modeling; optimization; VOLTAGE REGULATOR; OPTIMAL-DESIGN; CONVERTER; LOAD; COMPRESSION;
D O I
10.1145/2914791
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
The number of chip pins is limited due to the cost and reliability issues of sophisticated packages, and it is predicted that the chip pin count will be overstretched to satisfy the requirements of both power delivery and memory access. The gap between the achievable pin count and the demand will increase as the technology scales, due to the increasing computation resources and supply current. Pin reduction techniques are thus required for continued computing performance growth. In this article, we propose a chip pin constraint alleviation strategy, through on/off-chip power delivery system co-design, to effectively reduce the demand for power pins. An analytical model of a power delivery system, consisting of on/off-chip regulators and a power delivery network, is proposed to evaluate the influence of regulator design and package conduction loss. By combining this model with a multi-core processor model of performance and memory bandwidth requirements, we characterize the entiremulti-core processor system to investigate the relationship between the chip pin constraint and performance in multi-core processor scaling and the effectiveness of our strategy. Experiments show that with the conventional power delivery system design, the chip pin constraint severely limits the performance growth as the technology scales. Using the on/off-chip power delivery system codesign, our strategy achieves a significant pin count reduction, for example, 31.3% at the 8nm technology node, compared to the conventional design with the same chip performance, while, provided with the same chip pin count, it is able to improve, by 35.0%, the chip performance at 8nm compared to the conventional design. For real applications of different parallelism, our strategy outperforms its counterpart, with a 23.7% performance improvement on average at the 8nm technology node.
引用
收藏
页数:24
相关论文
共 38 条
[1]  
Alameldeen AR, 2004, CONF PROC INT SYMP C, P212
[2]  
[Anonymous], P ADV POW EL C
[3]  
[Anonymous], P 2014 29 ANN IEEE A
[4]  
Black B, 2006, INT SYMP MICROARCH, P469
[5]   Practical Strategies for Power-Efficient Computing Technologies [J].
Chang, Leland ;
Frank, David J. ;
Montoye, Robert K. ;
Koester, Steven J. ;
Ji, Brian L. ;
Coteus, Paul W. ;
Dennard, Robert H. ;
Haensch, Wilfried .
PROCEEDINGS OF THE IEEE, 2010, 98 (02) :215-236
[6]  
Esmaeilzadeh H, 2011, ISCA 2011: PROCEEDINGS OF THE 38TH ANNUAL INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE, P365, DOI 10.1145/2024723.2000108
[7]  
George V, 2007, IEEE ASIAN SOLID STA, P14
[8]  
Gupta MS, 2007, DES AUT TEST EUROPE, P624
[9]   Area-efficient linear regulator with ultra-fast load regulation [J].
Hazucha, P ;
Karnik, T ;
Bloechel, BA ;
Parsons, C ;
Finan, D ;
Borkar, S .
IEEE JOURNAL OF SOLID-STATE CIRCUITS, 2005, 40 (04) :933-940
[10]   A 233-MHz 80%-87% elfficient four-phase DC-DC converter utilizing air-core inductors on package [J].
Hazucha, P ;
Schrom, G ;
Hahn, J ;
Bloechel, BA ;
Hack, P ;
Dermer, GE ;
Narendra, S ;
Gardner, D ;
Karnik, T ;
De, V ;
Borkar, S .
IEEE JOURNAL OF SOLID-STATE CIRCUITS, 2005, 40 (04) :838-845