X10 on the single-chip cloud computer: Porting and preliminary performance

Keith Chapman; Ahmed Hussein; Antony L. Hosking

doi:10.1145/2212736.2212743

X10 on the single-chip cloud computer: Porting and preliminary performance

Keith Chapman^*, Ahmed Hussein, Antony L. Hosking

^*Corresponding author for this work

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

5 Citations (Scopus)

Abstract

The Single-Chip Cloud Computer (SCC) is an experimental processor created by Intel Labs. SCC is essentially a 'cluster-on-a-chip', so X10 with its support for places and remote asynchronous invocations is a natural fit for programming this platform. We report here on our experience porting X10 to the SCC, and show performance and scaling results for representative X10 benchmark applications. We compare results for our extensions to the SCC native messaging primitives in support of the X10 run-time, versus X10 on top of a prototype MPI API for SCC. The native SCC run-time exhibits better performance and scaling than the MPI binding. Scaling depends on the relative cost of computation versus communication in the workload used, since SCC is relatively underpowered for computation but has hardware support for message passing.

Original language	English
Title of host publication	Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11
DOIs	https://doi.org/10.1145/2212736.2212743
Publication status	Published - 2011
Externally published	Yes
Event	2011 ACM SIGPLAN X10 Workshop, X10 '11 - San Jose, CA, United States Duration: 4 Jun 2011 → 4 Jun 2011

Publication series

Name	Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11

Conference

Conference	2011 ACM SIGPLAN X10 Workshop, X10 '11
Country/Territory	United States
City	San Jose, CA
Period	4/06/11 → 4/06/11

Access to Document

10.1145/2212736.2212743

Cite this

@inproceedings{4ba81300af734df6926c26662f186053,

title = "X10 on the single-chip cloud computer: Porting and preliminary performance",

abstract = "The Single-Chip Cloud Computer (SCC) is an experimental processor created by Intel Labs. SCC is essentially a 'cluster-on-a-chip', so X10 with its support for places and remote asynchronous invocations is a natural fit for programming this platform. We report here on our experience porting X10 to the SCC, and show performance and scaling results for representative X10 benchmark applications. We compare results for our extensions to the SCC native messaging primitives in support of the X10 run-time, versus X10 on top of a prototype MPI API for SCC. The native SCC run-time exhibits better performance and scaling than the MPI binding. Scaling depends on the relative cost of computation versus communication in the workload used, since SCC is relatively underpowered for computation but has hardware support for message passing.",

author = "Keith Chapman and Ahmed Hussein and Hosking, {Antony L.}",

year = "2011",

doi = "10.1145/2212736.2212743",

language = "English",

isbn = "9781450307703",

series = "Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11",

booktitle = "Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11",

note = "2011 ACM SIGPLAN X10 Workshop, X10 '11 ; Conference date: 04-06-2011 Through 04-06-2011",

}

TY - GEN

T1 - X10 on the single-chip cloud computer

T2 - 2011 ACM SIGPLAN X10 Workshop, X10 '11

AU - Chapman, Keith

AU - Hussein, Ahmed

AU - Hosking, Antony L.

PY - 2011

Y1 - 2011

N2 - The Single-Chip Cloud Computer (SCC) is an experimental processor created by Intel Labs. SCC is essentially a 'cluster-on-a-chip', so X10 with its support for places and remote asynchronous invocations is a natural fit for programming this platform. We report here on our experience porting X10 to the SCC, and show performance and scaling results for representative X10 benchmark applications. We compare results for our extensions to the SCC native messaging primitives in support of the X10 run-time, versus X10 on top of a prototype MPI API for SCC. The native SCC run-time exhibits better performance and scaling than the MPI binding. Scaling depends on the relative cost of computation versus communication in the workload used, since SCC is relatively underpowered for computation but has hardware support for message passing.

AB - The Single-Chip Cloud Computer (SCC) is an experimental processor created by Intel Labs. SCC is essentially a 'cluster-on-a-chip', so X10 with its support for places and remote asynchronous invocations is a natural fit for programming this platform. We report here on our experience porting X10 to the SCC, and show performance and scaling results for representative X10 benchmark applications. We compare results for our extensions to the SCC native messaging primitives in support of the X10 run-time, versus X10 on top of a prototype MPI API for SCC. The native SCC run-time exhibits better performance and scaling than the MPI binding. Scaling depends on the relative cost of computation versus communication in the workload used, since SCC is relatively underpowered for computation but has hardware support for message passing.

UR - http://www.scopus.com/inward/record.url?scp=84862624230&partnerID=8YFLogxK

U2 - 10.1145/2212736.2212743

DO - 10.1145/2212736.2212743

M3 - Conference contribution

SN - 9781450307703

T3 - Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11

BT - Proceedings of the 2011 ACM SIGPLAN X10 Workshop, X10 '11

Y2 - 4 June 2011 through 4 June 2011

ER -

X10 on the single-chip cloud computer: Porting and preliminary performance

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this