18-24 June 2017
Palacio de Congresos
Europe/Madrid timezone
Contribution Parallel
Seminarios 8
Software Development
Performance Portability Strategies for Grid C++ Expression Template
Speakers
- Dr. Meifeng LIN
Primary authors
- Prof. Peter BOYLE (University of Edinburgh)
- Dr. Kate CLARK (NVIDIA)
- Prof. Carleton DETAR (University of Utah)
- Dr. Meifeng LIN (Brookhaven National Laboratory)
- Dr. Verinder RANA (Brookhaven National Laboratory)
- Dr. Alejandro VAQUERO AVILÉS-CASCO (University Of Utah)
Files
Content
One of the key requirements for the Lattice QCD Application Development as part of the US Exascale Computing Project is performance portability across multiple architectures. Using the Grid C++ expression template as a starting point, we report on the progress made with regards to the Grid GPU offloading strategies. We present both the successes and issues encountered in using CUDA, OpenACC and Just-In-Time compilation. Experimentation and performance on GPUs with a SU(3)×SU(3) streaming test will be reported. We will also report on the challenges of using current OpenMP 4.x for GPU offloading in the same code.
Preferred track (if multiple tracks have been selected)
Software Development
