Accelerate routines within the CCSM4 project using GP-GPU
Location: Niels Bohr Institutet, University of Copenhagen (UCHP), Copenhagen, Denmark
Project Abstract:
The Community Earth System Model (CESM) is a fully-coupled, global climate model that provides state-of-the-art computer simulations of the Earth’s past, present, and future climate states. In order to improve the simulations, the resolution and complexity of the model is constantly increased, which in turn increases the computational demand of the model.
The proposed project is to use GP-GPU in order to answer this increasingly computational demand. The student will analyse CCSM4 and identify functions that may benefit from GP-GPU execution and then port these functions to GP-GPU.
Global projections of surface temperature monthly mean for January 2099
Project Mentor: Mads Kristensen
Site Co-ordinator: Brian Vinter
Learning Outcomes:
The student will learn how to:
- Identify code that will benefit from GP-GPU execution
- Use OpenCL in order to utilize GP-GPUs
- Implement a subset of a climate model simulation
Student Prerequisites (Compulsory):
- Basic linux skills
- Elementary knowledge of parallel computing
Student Prerequisites (Desirable):
- Experience with OpenCL
Training Materials:
http://developer.amd.com/tools/heterogeneous-computing/amd-accelerated-parallel-processing-app-sdk/
http://www.cesm.ucar.edu
Project Application Reference: Denmark – UHCP – Accelerate Routines in CCSM4
Applications are now closed
