NeuroGPU: GPU Accelerated Neuronal Modeling
The project involved convex optimization and GPU programming to speed up neuronal simulations. Essentially, given the description of a neuron and/or a neuronal network, the project aims to translate this high level description into something that can be run on a GPU. GPU-bound simulations are on the order of 1000’s of times faster than CPU-bound simulations. This summer, I worked on porting these simulations to Python, where they can be more easily accessed by programmers and streamlining the efficiency of the modeling program.
Message To SponsorThis research has meant a lot to me. I have spent a lot of time working on this project and have learned a lot not only about GPU programming, but also Python programming as well. I have developed significant portions of the Python code and have made important design decisions that will hopefully help future programmers using this program and future researchers who help contribute to this project. Additionally, I have also learned to work in a research setting and have gotten used to the research process, something which has been an invaluable experience for me.
Mentor: Dr. Roy Ben-Shalom/Dr. Kevin Bender | Neuronscience (UCSF)