Joss Knight

Last seen: 1 day ago

MathWorks

380 total contributions since 2013

Although I cannot be contacted directly, if you would like to ask me a question all you have to do is mention "GPU" somewhere in your MATLAB Answers question.

Joss Knight's Badges

View details...

Answered
Where does the file"mex_CUDA_win64.xml " exist in Matlab2020a?
The correct method for compiling GPU MEX functions for many years now has been to use the utility mexcuda, which does not requir...

3 months ago | 0

Answered
How do I substitute all the activation functions of a neural network?
It looks as though you've replaced every relu layer with multiple other layers. This will make your network deeper. The deeper t...

3 months ago | 0

Answered
Using mex with CUDA
Hi Jack. I expect the answer to this question is that you have installed a more recent version of the CUDA toolkit, specifically...

3 months ago | 0

Answered
How to run a custom function to gpu?
I guess the basic problem is that the gctest function doesn't support gpuArray data. You could try taking a copy of it and fixin...

3 months ago | 0

Answered
How do I take poisson draws with poissrnd() using GPU resources via arrayfun?
As you can see from the documentation, poissrnd is not a function that is support by the FUN argument of gpuArray/arrayfun. You ...

4 months ago | 0

Answered
Same gpu operation in loop but two speeds
MATLAB will do some optimisations when it sees you are doing the same thing repeatedly. In your case, the optimisation that appl...

6 months ago | 0

| accepted

Answered
Parallrl calculations for Deep learning Toolbox
Sorry about this not-very-good error, which should be fixed in the current release. What it means is that 'Shuffle', 'never' is ...

6 months ago | 1

Answered
Error using parpool (line 113)
This is a very unusual error and indicates a problem with your installation. I suspect your MATLAB path is mixed up with previou...

6 months ago | 2

Answered
CPU/GPU Architecture Recommendation
For Deep Learning the RTX 2080 is a good choice.

6 months ago | 0

Answered
Using external GPU with MatLab 2020a
There may be nothing you can do. Data transfer times to an external GPU are much higher. It might help to increase the MiniBatch...

7 months ago | 0

Answered
Arrayfun GPU in "Game of Life" works slower than CPU
Check out this Answer. The arrayfun version is rather dependent on good memory performance since the kernel is accessing global ...

7 months ago | 0

| accepted

Answered
Deep Learning - Distributed GPU Memory
No, there is nothing like what you are after, to distribute the weights of a fully connected layer across multiple GPUs. You cou...

7 months ago | 0

Answered
Multiple GPUs perform slower than single GPU to train a semantic segmentation network
On Windows, due to GPU communication issues on that platform, it is difficult to get any benefit from multi-GPU training. This w...

8 months ago | 0

| accepted

Answered
The utilization of GPU is low in deep learning
Try following some of the advice in the following MATLAB Answer: https://uk.mathworks.com/matlabcentral/answers/463367-gpu-utili...

8 months ago | 0

Answered
Assigning gpuArrays to different graphics cards
There is no way to do what you ask. Selecting a GPU is the only way to move data there, and selecting a GPU resets all GPU data....

8 months ago | 0

| accepted

Answered
GPU out of memory
In your example code you are using the default mini-batch size of 128. Reduce the MiniBatchSize training option until you stop g...

9 months ago | 2

Answered
incorrect memory copy when using property size validator
I was able to reproduce this in R2018b but not in R2019a or R2019b. It looks like property validators used to trigger a deep cop...

9 months ago | 0

| accepted

Answered
Solution of large sparse matrix systems using GPU MLDIVIDE
The general advice is that Sparse MLDIVIDE may be convenient, but it is 'usually' slower than use of an iterative solver with an...

9 months ago | 0

| accepted

Answered
Deep Learning: Training Network with "parallel" option using only CPUs
Even with a weak graphics card you will usually see better performance than on multiple CPUs. However, to try it out, after you ...

10 months ago | 0

| accepted

Answered
How to use Levenberg-Marquardt backprop with GPU?
This isn't supported out of the box yet. You could convert your network to use dlarray and train it with a custom training loop....

10 months ago | 0

Answered
hardware requierments for MATLAB
Partial answer GPU Computing. You can't have a MATLAB without CPU Computing so obviously both is better. No Mostly using the ...

11 months ago | 1

Answered
Which Visual Studio 2019 package should I install to work with CUDA?
To accelerate your MATLAB code with an NVIDIA GPU, you do not need to install a C++ Compiler.

11 months ago | 0

Answered
Why would the file size of a deep learning gradient become much bigger after saving as a .mat file?
The difference is that whos is unable to account for the fact that the data is all stored on the GPU, and is only showing CPU me...

11 months ago | 0

| accepted

Answered
Training a Variational Autoencoder (VAE) on sine waves
It looks like your input data size is wrong. Your formatting says that the 4th dimension is the batch dimension, but actually it...

11 months ago | 0

| accepted

Answered
Is lhsdesign (latin hypercube sampling) supported by gpuArray?
It is not supported. You can tell whether or not a function supports gpuArray, more reliably than from the list of gpuArray meth...

12 months ago | 1

| accepted

Answered
DOES GEFORCE GTX1080 GPU WORKS WELL FOR DEEP LEARNING TRAINING??
Yes.

12 months ago | 1

| accepted

Answered
Feed data into Neural Networks file-by-file
Datastores are designed for precisely this purpose. It may be that you're after an imageDatastore processed by a transform.

12 months ago | 0

Answered
Gather cell array from GPU to CPU
A_cpu = cellfun(@gather, A_gpu, 'UniformOutput', false);

12 months ago | 1

| accepted

Answered
Error in matlab included deep learning example
There is a bug in this Example which will be rectified. Thanks for reporting. To workaround, initialize the loss variable in the...

1 year ago | 2

| accepted

Answered
movsum slower than conv2 in GPU
One might theorize, perhaps, that movsum literally uses the same kernels as conv2, but first has to construct the filter of ones...

1 year ago | 2