Loading OpenMP to Cell: An Effective Compiler Framework for Heterogeneous Multi-core Chip

Authors:
Haitao Wei;Junqing Yu
Affiliations:
College of Computer Science & Technology, Huazhong University of Science & Technology, Wuhan, China 430074;College of Computer Science & Technology, Huazhong University of Science & Technology, Wuhan, China 430074
Venue:
IWOMP '07 Proceedings of the 3rd international workshop on OpenMP: A Practical Programming Model for the Multi-Core Era
Year:
2007

Citing 2
Cited 1

Using advanced compiler technology to exploit the performance of the Cell Broadband EngineTM architecture

IBM Systems Journal
CellSs: a programming model for the cell BE architecture

Proceedings of the 2006 ACM/IEEE conference on Supercomputing

Video processing and retrieval on cell processor architecture

ICEC'07 Proceedings of the 6th international conference on Entertainment Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The Cell employs a multi-core-on-chip design that integrates 9 processing elements, local memory and communication hardware into one chip. We present a source to source OpenMP compiler framework which translates the program with OpenMP directives to the necessary codes for PPE and SPE to exploit the parallelism of a sequential program through the different processing elements of the Cell. Some effective mapping strategies are also presented to conduct the thread creating and data handling between the different processors and reduce the overhead of system performance. The experimental results show that such compiler framework and mapping strategy can be effective for the heterogeneous multi-core architecture.