Answered
GPUを使用したDeepLearningの並列処理に関する質問
MATLAB CNNs support any modern multi-gpu environment and it's difficult to advise. Amazon and Azure provide cloud environments o...

8 years ago | 1

Answered
gpuArray/filter doesn't support multi-channel lag terms?
This is a bug that was fixed in R2016b, so you must have an earlier version of MATLAB. If you can't upgrade, you can avoid it by...

8 years ago | 0

| accepted

Answered
Are there any GPU which is compatible both with MATLAB and a MacBook?
An older line of MacBook Pro has the NVIDIA GTX 750M which is compatible with MATLAB. They don't make them any more but you migh...

8 years ago | 0

Answered
How to change activation function for fully connected layer in convolutional neural network?
Activations are added as a separate layer, and in R2017a there is only the |RelULayer| (see <https://www.mathworks.com/help/nnet...

8 years ago | 4

| accepted

Answered
Limit total GPU memory usage
Since several releases ago MATLAB releases variables as soon as they go out of scope, are overwritten, cleared (using |clear| or...

9 years ago | 1

Answered
Why for loop in CPU faster than in GPU
This isn't amazing at all. You are launching 7 million kernels, each to copy one value into GPU memory. The kernels cannot overl...

9 years ago | 0

Answered
GPU for ConvNN training out of memory
You're going to struggle massively with a small mobile chip like the 750M with only 2GB of memory and a large network like AlexN...

9 years ago | 0

Answered
Cubic interp1 on a GPU array: bug?
This looks like a straightforward bug in the error message that's come as a result of a change in the default |'cubic'| algorith...

9 years ago | 0

| accepted

Answered
Unable to compile cuda code containing dynamic parallelism: Error: “ptxas fatal : Unresolved extern function 'cublasCreate_v2'”
You need to link against the cublas device library in the device linking stage and unfortunately there isn't a proper formal API...

9 years ago | 0

| accepted

Answered
Unable to compile cuda code containing dynamic parallelism: Error: “ptxas fatal : Unresolved extern function 'cublasCreate_v2'”
I don't think you can call cublasDgemm inside a kernel - that's not the way dynamic parallelism works. You can't just call any o...

9 years ago | 0

Answered
Can I use cpu instead of gpu for the DeepLearningImageClassificationExample.m
<https://www.mathworks.com/help/nnet/ref/seriesnetwork.activations.html#namevaluepairs Yes>

9 years ago | 0

Answered
Matlab bug and deadly error while trying to visualize FasterRCNN network feature maps from softmax layer
The |activations| function of the |vision.cnn.FastRCNN| is undocumented and Hidden and so not officially supported. Its inputs a...

9 years ago | 0

Answered
CUDAKernel/feval error, cannot find 'single' input.
I presume if you ask classUnderlying(patches) it'll probably tell you they are single. After all, you have just declare...

9 years ago | 1

| accepted

Answered
problem with using GPU in matlab
The GTX 1050 doesn't have very good double precision performance, and the default precision for all MATLAB computation is double...

9 years ago | 0

| accepted

Answered
How to utilize power of GPU computing in Matlab
<https://www.mathworks.com/products/parallel-computing.html Start here.>

9 years ago | 0

Answered
run gpu function using matlab
The argument parser cannot handle your type qualifier 'volatile' in order to work out the appropriate datatypes of arrays to pas...

9 years ago | 0

Answered
Preallocation of composites using smpd
It seems like this is just an issue of timing and synchronisation. You can see this by adding a call to |wait(g)| at the end of ...

9 years ago | 1

| accepted

Answered
why two different mini-batch Accuracy in CNN
You need to <https://www.mathworks.com/support/bugreports/1439741 install the patch>.

9 years ago | 0

Answered
Unable to get less time for PARFOR loop than FOR loop on the system which has Tesla K20c GPU card in it .
You can try using <https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiIp8Gro-3TAhUGJMA...

9 years ago | 0

Answered
Generate C/C ++ source code with maximum optimization
Nope, sorry. The third one is supported for some kinds of codegen.

9 years ago | 0

Answered
Better performace: GPU + Workers
It depends what you're optimizing. Use of the GPU only really make sense if the objective function is a large enough operation t...

9 years ago | 0

| accepted

Answered
Evaluating fucntion by GPU.
Your question is extremely confusing. Every function can act on every element of its inputs at once. So if I write a function ...

9 years ago | 0

Answered
how to use parallel computing with train faster rcnn detector.
The Error says that |'multi-gpu' and 'parallel'| are not supported and yet you have tried to set your |'ExecutionEnvironment'| t...

9 years ago | 0

Answered
How do I compile cuda library dll files so that I can use them from Matlab.
PTX is device code only so it's only the runtime CUDA libraries you can link in. You need to learn more about CUDA device code g...

9 years ago | 0

Answered
Will Matlab support AMD GPU in future?
We periodically review support for OpenCL and AMD. At the moment it doesn't quite have sufficient support to provide the integra...

9 years ago | 3

Answered
How to distribute computation on GPU vector-wise?
You can loop over and read multiple entries in an input array (as an up-value variable) inside |arrayfun|, but you can't loop ov...

9 years ago | 0

Answered
How to convert this vectorized code into GPU code for MAXIMUM speedup ?
I was able to get a marginal speedup with additional vectorization of the mask: x = sum(I < cat(3, 120, 155, 160), 3) == t...

9 years ago | 0

Answered
GPU requirement for convolution neural network.
Current best value for deep learning, image processing and computer vision applications is the GeForce GTX 1080i or the Titan X ...

9 years ago | 0

| accepted

Answered
indexing sparse gpuArrays?
Generally if you are indexing scalar values in a sparse gpuArray then you're trying to do something very inefficient. Still, if ...

9 years ago | 0

| accepted

Answered
How gpuArray.randperm is implemented
|randperm| calls [~,p] = sort(rand(1,N)) but obviously does everything on the GPU.

9 years ago | 0

| accepted