Hybrid CPU-GPU computing for simulating calcium handling in the heart
The overall objective is to develop a precise understanding of the role played by disturbances of calcium handling in cardiac arrhythmias. In this master project, computer simulations on cutting-edge hardware (especially on clusters that are made up both both multicore CPUs and general-purpose GPUs) will be chosen as the tool. The candidate will implement high-efficiency simulators that are capable of hybrid CPU-GPU computnig, and also carry out physiologically realistic simulations.
Example results of a simulation of calcium handling
What you will do:
Start from an exsiting CPU code, implement a simulator that is able to use both CPUs and GPUs
Optimize the perfromance of the hybrid simulator
Carry out actual simulations under the guidance of experts
Visualize and analyze simulation results
What you will learn:
Standard parallel programming techniques for CPUs (such as MPI and OpenMP)
CUDA programming that is the de-facto standard for GPUs
Important concepts of computational science
Basic knowledge about calcium handling in the heart
