Performance, Power Efficiency and Scalability of Asymmetric Cluster Chip Multiprocessors

IEEE Computer Architecture LettersPublished 1 January 2006

Tomer Y. Morad, Uri Weiser, Avinoam Kolodny, Mateo Valero, Eduard Ayguadé

DOI10.1109/l-ca.2006.6

Source URL

Citations155

SJR quartileQ2

SJR score0.50

SNIP0.56

TL;DR

Evaluated ACCMP architectures as a mechanism to achieve the highest performance for a given power budget show that asymmetric multiprocessors can reduce power consumption by more than two thirds with similar performance compared to symmetric multipROcessors.

Abstract

This paper evaluates asymmetric cluster chip multiprocessor (ACCMP) architectures as a mechanism to achieve the highest performance for a given power budget. ACCMPs execute serial phases of multithreaded programs on large high-performance cores whereas parallel phases are executed on a mix of large and many small simple cores. Theoretical analysis reveals a performance upper bound for symmetric multiprocessors, which is surpassed by asymmetric configurations at certain power ranges. Our emulations show that asymmetric multiprocessors can reduce power consumption by more than two thirds with similar performance compared to symmetric multiprocessors

Keywords

Computer Science

Validity of the single processor approach to achieving large scale computing capabilities

4,200 Citations1967Gene M. Amdahl

ACM SIGARCH Computer Architecture NewsSingle-ISA Heterogeneous Multi-Core Architectures for Multithreaded Workload Performance

521 Citations2004Rakesh Kumar, Dean M. Tullsen +3 more

This paper examines two single-ISA heterogeneous multi-core architectures in detail, demonstrating dynamic core assignment policies that provide significant performance gains over naive assignment, and even outperform the best static assignment.

ComputerA single-chip multiprocessor

409 Citations1997Basem A. Nayfeh, Kunle Olukotun

Presents the case for billion-transistor processor architectures that will consist of chip multiprocessors (CMPs): multiple (four to 16) simple, fast processors on one chip, and all processors share a larger level-two cache.

Single-ISA heterogeneous multi-core architectures for multithreaded workload performance

250 Citations2004Rakesh Kumar, Dean M. Tullsen +3 more

Lecture notes in computer scienceSPEComp: A New Benchmark Suite for Measuring Parallel Computer Performance

235 Citations2001Vishal Aslot, Max Domeika +4 more

An overview of a new benchmark suite for parallel computers, SPEComp, which targets mid-size parallel servers and includes a number of science/engineering and data processing applications, is presented.

ACM SIGARCH Computer Architecture NewsMitigating Amdahl's Law through EPI Throttling

200 Citations2005Murali Annavaram, Ed Grochowski +1 more

Best of both latency and throughput

137 Citations2004Ed Grochowski, Ronny Ronen +2 more

This paper surveys four techniques for achieving variable energy per instruction and concludes that a combination of asymmetric cores and voltage/frequency scaling offers the most promising approach to design a chip-level multiprocessor that can achieve both excellent latency performance and excellent throughput performance.

Parallel ComputingPerformance of parallel processors

106 Citations1989Horace P. Flatt, Ken Kennedy

Bounds are derived under fairly general conditions on the synchronization cost function that have implications for a variety of parallel architecture and can be used to derive several popular ‘laws’ about processor performance and efficiency.

Cost-performance analysis of heterogeneity in supercomputer architectures

34 Citations2002Daniel A. Menascé, Virgı́lio Almeida

Heterogeneity in supercomputer architectures is a quite promising design approach that deserves further investigation, according to Grosch's law, which states that computer performance increases as the square of its cost.

QueueGetting Gigascale Chips

27 Citations2003Shekhar Borkar

Advances in software technology, such as rich multimedia applications and runtime systems, exploited this performance explosion, delivering to end users higher productivity, seamless Internet connectivity, and even multimedia and entertainment.