OpenCL code generation for low energy wide SIMD architectures with explicit datapath.

D. She, Y. He, L.J.W. Waeijen, H. Corporaal

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

6 Citations (Scopus)
1 Downloads (Pure)


Energy efficiency is one of the most important aspects in designing embedded processors. The use of a wide SIMD processor architecture is a promising approach to build energy-efficient high performance embedded processors. In this paper, we propose a configurable wide SIMD architecture that utilizes explicit datapath to achieve high energy efficiency. To efficiently program the proposed architecture with a standard parallel programming language, we introduce a tool flow that can compile and map OpenCL programs onto it. The compiler in the proposed tool flow is able to analyze the static access patterns in OpenCL kernels and generate efficient mapping and code that utilizes the explicit datapath. Experimental results show that the proposed architecture is efficient. In a 128-PE processor, the proposed architecture is able to achieve over 200 times speed-up and reduce the energy consumption of register file and memory by over 90% compared to a RISC processor.
Original languageEnglish
Title of host publicationProceedings of the 13th International Conference on Embedded Computer Systems : Architectures, Modeling and Simulation (SAMOS XIII)
EditorsH. Jeschke, O. Silvén
Publication statusPublished - 2013
Eventconference; SAMOS 2013; 2013-07-15; 2013-07-18 -
Duration: 15 Jul 201318 Jul 2013


Conferenceconference; SAMOS 2013; 2013-07-15; 2013-07-18
OtherSAMOS 2013


Dive into the research topics of 'OpenCL code generation for low energy wide SIMD architectures with explicit datapath.'. Together they form a unique fingerprint.

Cite this