I am currently using Matlab 2019b version. How do I enable the GPU on my laptop for certain image processing applications like Fuzzy Connected adaptive segmentation(Not deep learning where we get a promt if we want to switch from CPU to the GPU)?

The system is becoming too slow when I tried to execute the Fuzzy connected adaptive segmentation algorithm and it kept on executing without an output for almost 8hrs at a stretch for an image size of approx. 1024x1024.

What are the changes to be made in the code ( such as using specific keywords) to ensure that the code executes on the GPU?

I am using a Dell laptop, i7 and the hardware configuration says there are two GPUs? How do I enable them?

Thanks in advance

Answer 1

[Manish Mahalwal]

Hi Sushma,

As the first step, to run your code on the GPU, you need to have the Parallel Computing Toolbox installed.

You need to use the gpuArray to transfer your data from CPU to GPU. Now, not all the image processing functions accept data from the GPU. The list of the supported functions is available here.

Thirdly, to transfer data back from the GPU to CPU, you can use the gather function.

The results may slightly vary from what you would have got when you execute the same code on CPU completely.

Some more info here: https://www.mathworks.com/help/images/image-processing-on-a-gpu.html

DISCLAIMER: These are my own views and in no way depict those of MathWorks.

[Sushma TV]

Thank you for the answer!! I will check out and see if I am able to progress with the inputs given.

Thanks again.

[Jason Ross]

Note that you also need a supported nVidia GPU and nVidia driver installed on your laptop.

[Sushma TV]

Thank you!!

[Sushma TV]

Further on the same doubt of being able to use the GPU on my laptop..I have an AMD Radeon R7 M270 GPU..Is it possible to use this on Matlab?

[Manish Mahalwal]

Unfortunately AMD GPUs cannot be used to manipulate gpuArray objects. Like Jason mentioned only nVidia GPUs are supported. The types supported by MATLAB are available here.