How to use GPU with your matlab program?

Recently I purchased a laptop (with GPU) so that my matlab programs can run faster. However, I am unaware of the process of utilizing GPU in a program? is that even a thing?

Do I need to used some special functions or acticate GPU in my program so that the GPU is utilized properly?

Answer 1

[KSSV]

To run your code in GPU, you need not to do much. It is very easy and straight forward. You need to place the variable/ data into GPU using the function gpuarray . Read here: https://in.mathworks.com/help/parallel-computing/gpuarray.html

If the variable is of class GPU, it will be placed in GPU and all the operations you do will be on GPU. Most of the functions support gpu calculations. Find it here: https://in.mathworks.com/help/referencelist.html?type=function&capability=gpucodegen

Example:

% Multiplication of matrices in GPU 
A = rand(100) ; 
B = rand(100) ; 
C = A*B ;  % runs on CPU 
%
A_gpu = gpuarray(A) ; 
B_gpu = gpuarray(B) ;
C_gpu = A*B ;   % runs on GPU 

Running on GPU is easy and straight forward. Read here: https://in.mathworks.com/help/parallel-computing/run-matlab-functions-on-a-gpu.html

[Virtualkeeda]

oh ok. Thanks, I will update my program accordingly.

I more thing, So if I want to run the program on a server (cluster or a HPC), then again I just need to use this gpuArray and it will work??

[Virtualkeeda]

Hello @KSSV,

I have computations on structs in my code with a label identifier and other numerical values. What I could understand is I can use gpuArray only for numerical computations but whenever it encounter a local operation, it doesnt work. Is there a way out of this problem?

I am trying to run code available at: https://in.mathworks.com/matlabcentral/fileexchange/42853-deep-neural-network

Can you please help?