Abstract
In this work, we evaluate several emerging parallel programming models: Kokkos, RAJA, OpenACC, and OpenMP 4.0, against the mature CUDA and OpenCL APIs. Each model has been used to port Tealeaf, a miniature proxy application, or mini app, that solves the heat conduction equation and belongs to the Mantevo Project. We find that the best performance is achieved with architecture-specific implementations but that, in many cases, the performance portable models are able to solve the same problems to within a 5% to 30% performance penalty. While the models expose varying levels of complexity to the developer, they all achieve reasonable performance with this application. As such, if this small performance penalty is permissible for a problem domain, we believe that productivity and development complexity can be considered the major differentiators when choosing a modern parallel programming model to develop applications like Tealeaf.
Original language | English |
---|---|
Article number | e4117 |
Number of pages | 20 |
Journal | Concurrency and Computation: Practice and Experience |
Volume | 29 |
Issue number | 15 |
Early online date | 29 Mar 2017 |
DOIs | |
Publication status | Published - 10 Aug 2017 |
Keywords
- Kokkos
- OpenMP 4.0
- performance portability
- programming models
- RAJA
Fingerprint
Dive into the research topics of 'Assessing the performance portability of modern parallel programming models using TeaLeaf'. Together they form a unique fingerprint.Profiles
-
Professor Simon N McIntosh-Smith
- School of Computer Science - Professor in High Performance Computing
- Microelectronics
Person: Academic , Group lead