Prototyping Algorithms and Testing CUDA Kernels in MATLAB

Originally published at:

This guest post by Daniel Armyr and Dan Doherty from MathWorks describes how you can use MATLAB to support your development of CUDA C and C++ kernels. You will need MATLAB, Parallel Computing Toolbox™, and Image Processing Toolbox™ to run the code. You can request a trial of these products. For a more detailed description of this…

Dear Mark. Although there are an abundance of information about Cudo computing with Matlab, it fails to mention which cards will work with Matlab. Matlab does not support all cards so one has to be careful. I'm interested in the K6000 to sue with Matlab, but I'm not sure whether its compatible since from other sources you read that its not added . See here for example:
They will only mention Java, Python etc.. So will the k6000 work with Matlab please?


Hi dan, thanks for your comment! The answer is yes. MATLAB Parallel
Computing Toolbox supports GPUs of compute capability 1.3 and higher.
Formal Kepler support was added a version or two ago. See

K6000 should work. This page says that GPUs with Compute Capability 2.0 or later are supported.