GSoC 2022 Project Idea 1.2: Improving GPU-accelerated Monte Carlo simulations in Disimpy: Code refactoring and optimization (350 h)

Disimpy is a GPU-accelerated diffusion-weighted magnetic resonance simulator that is useful in the development and validation of neuroimaging methods. Disimpy is written in Python, making the source code very approachable to researchers with little or no prior experience in GPU-computing.

Project: Code refactoring and optimization to reduce simulation runtimes and GPU memory usage.

Skill level: Advanced

Required skills: Python, Numba, CUDA.

Time commitment: Full-time

Lead mentor: Leevi Kerkelä @kerkelae

Project website:

Backup mentors: Rafael Neto Henriques, Marco Palombo

Tech keywords: Python, NumPy, Numba, CUDA, 3D modelling

Hi malin,
I’m greatly interested in this project and have some related knowledge. But I’m wondering how to join your project in GSoC 2022? Is there any e-mail address of mentor I can reach to?

You can send me an email (

Hello, can you send me a message? I’m interested too :smiley:

Hi, happy to help you by answering any questions you may have regarding Disimpy.