2005
DOI: 10.1147/rd.492.0195
|View full text |Cite
|
Sign up to set email alerts
|

Overview of the Blue Gene/L system architecture

Abstract: The Blue Genet/L computer is a massively parallel supercomputer based on IBM system-on-a-chip technology. It is designed to scale to 65,536 dual-processor nodes, with a peak performance of 360 teraflops. This paper describes the project objectives and provides an overview of the system architecture that resulted. We discuss our application-based approach and rationale for a low-power, highly integrated design. The key architectural features of Blue Gene/L are introduced in this paper: the link chip component a… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
164
0
1

Year Published

2008
2008
2012
2012

Publication Types

Select...
5
3
2

Relationship

0
10

Authors

Journals

citations
Cited by 290 publications
(170 citation statements)
references
References 22 publications
0
164
0
1
Order By: Relevance
“…The Blue Gene/P system design [1] is very similar to Blue Gene/L [2][3][4]. Blue Gene compute nodes (CNs) run the user application on a lightweight compute node kernel (CNK), and do not have direct TCP/IP connection to an external network.…”
Section: Gpfs Architecture On Blue Genementioning
confidence: 99%
“…The Blue Gene/P system design [1] is very similar to Blue Gene/L [2][3][4]. Blue Gene compute nodes (CNs) run the user application on a lightweight compute node kernel (CNK), and do not have direct TCP/IP connection to an external network.…”
Section: Gpfs Architecture On Blue Genementioning
confidence: 99%
“…It enables the application to take advantage of optimized implementations of collective operations. Several architectures, such as BlueGene [13], or networks, such as Quadrics [14], support hardware-optimized collective operations (they even perform some simple reduction operations directly in the network hardware). This implementation inherently supports the problematic case where the number of keys is small (the parallelism in the reduce phase is limited).…”
Section: Scalable Mapreduce In Mpimentioning
confidence: 99%
“…However, today's high performance computers consume tremendous amounts of energy. For example, a 360-Tflops supercomputer (such as IBM Blue Gene/L) with conventional processors requires 20 MW to operate, which is approximately equal to the sum of 22,000 US households power consumption [1], [2]. Furthermore, it is estimated that servers consume 0.5 percent of the world's total electricity usage [3], which if current demand continues, is projected to quadruple by 2020.…”
Section: Introductionmentioning
confidence: 99%