MATLAB: Fmincon GPU Cuda Acceleration using Parallel Computing Toolbox

Question

Can I use fmincon with GPU CUDA acceleration using Parallel Computing Toolbox? Becouse fmincon in not mentioned in the list of functions here: http://www.mathworks.com/help/distcomp/run-built-in-functions-on-a-gpu.html

Answer 1

The way you would exploit the GPU in fmincon would be to accelerate the computation of your objective function and its derivatives. To do that, you just use gpuArray objects for heavy and highly parallel computations inside the objective function, just like you would in any other code.

On the other hand, if the computational cost of each objective function evaluation is low, parallel computing is probably not going to be of any advantage.