I’m trying to assess whether Julia can be used for my use case. I have a ~25 dimensional differential equation whose solutions I’d to analyze for a variety of initial conditions and choices of parameters. As such, I have quite a few simulations to run. Is there a way to use Julia on GPUs to parallelize this task?
Yes there is a way, I don’t the exact details but I believe GitHub - SciML/DiffEqGPU.jl: GPU-acceleration routines for DifferentialEquations.jl and the broader SciML scientific machine learning ecosystem is the way to do it :). Also here for some examples Parallel Ensemble Simulations · DifferentialEquations.jl