mehr als 11 Jahre vor | 1

| akzeptiert

Beantwortet
decimal to binary conversion of gpuArray gives error
Unfortunately, as you have discovered, the |gpuArray| version of |log2| doesn't support the second output argument as needed by ...

mehr als 11 Jahre vor | 0

Beantwortet
can't do parallel computing
You might need to clear out the "Storage Location" directory on disk - you can find this by executing: c = parcluster('loca...

mehr als 11 Jahre vor | 6

Beantwortet
Implementing batched mldivide on the GPU
This is now supported in R2014b - see <http://www.mathworks.com/help/distcomp/release-notes.html the release notes>.

mehr als 11 Jahre vor | 0

Beantwortet
Speed up a program
Here's what I tried with PARFOR, and this worked: function D = pfeg N = 100; D = zeros(N, N, 3); parfor i = 1:N for...

mehr als 11 Jahre vor | 0

Beantwortet
How to optimize gpuArray operation to minimize GPU memory
Your card cannot possibly fit 3 1.6Gb matrices in memory, which is needed to perform |C = A - B|. Therefore, you need to make yo...

mehr als 11 Jahre vor | 0

Beantwortet
Problems in sharing mex codes with workers in a cluster
You need to use matlabpool('addfiledependencies', {'mex1', 'mex2'}) to make the MEX files available to the workers. See...

mehr als 11 Jahre vor | 0

| akzeptiert

Beantwortet
Parfor variable slicing and assignment
MATLAB uses a language analysis to work out what's going on the body of your |PARFOR| loop. In this case, you would like |A| to ...

mehr als 11 Jahre vor | 1

Beantwortet
Parpool local too slow
Unfortunately there is no way around this, other than to open a |parpool| when you first need one (or let it open automatically ...

mehr als 11 Jahre vor | 0

| akzeptiert

Beantwortet
Performance of selecting a subset of a GPUArray
Logical indexing on the GPU is slower than indexing with doubles. So, of you reuse a logical indexing vector, you may well find ...

mehr als 11 Jahre vor | 0

Beantwortet
GPU parallel computing on Intel
Unfortunately, Parallel Computing Toolbox supports only Nvidia GPUs. <http://www.mathworks.com/products/availability/#DM>

mehr als 11 Jahre vor | 0

Beantwortet
How to use multiple GPUs asynchronously
If you have 2 GPUs, you could use |SPMD| like this: parpool('local', 2); % ensure the pool is of size 2 spmd % set ...

mehr als 11 Jahre vor | 1

| akzeptiert

Beantwortet
r2014b cannot start parallel pool
This looks to me like a problem with licensing. Please contact install support to resolve this problem.

mehr als 11 Jahre vor | 0

| akzeptiert

Beantwortet
GPU arrayfun with shared arrays
Can you give a more concrete example of what you'd like to do with |A|, |B|, and |C|? You might be able to use a nested function...

mehr als 11 Jahre vor | 0

Beantwortet
Is it possible to conditionally switch between parallel and non parallel for loops?
You can do this by setting the optional "number of workers" argument to |0|. For example runInSerial = <...>; if runInSe...

mehr als 11 Jahre vor | 6

| akzeptiert

Beantwortet
How to use arrayfun for nested loops without using memory multiple times for the same read only data
Use |bsxfun| bsxfun(@(x,y) x + y, gpuArray(1:4), gpuArray(1:4).')

mehr als 11 Jahre vor | 0

| akzeptiert

Beantwortet
error with arrayfun and GPU computing part 2.
When working with |gpuArray| data, the function that you pass to the |arrayfun| will be called with scalar values. In other word...

mehr als 11 Jahre vor | 2

| akzeptiert

Beantwortet
Minimize MDCS Data Transfer
You could use the <http://www.mathworks.com/matlabcentral/fileexchange/31972-worker-object-wrapper Worker Object Wrapper> which ...

mehr als 11 Jahre vor | 0

| akzeptiert

Beantwortet
The variable g in a parfor cannot be classified..
It looks like you could simply replace |g(i)| with |g| and then hoist the inner for-loop bounds calculation out of the |PARFOR| ...

fast 12 Jahre vor | 0

Beantwortet
Help with multilevel parallel programming
One of the reasons that PARFOR does not support nested parallelism is that it is difficult to achieve good load-balancing in thi...

fast 12 Jahre vor | 0

Beantwortet
parallel perfix sum in matlab
In MATLAB, <http://www.mathworks.com/help/matlab/ref/cumsum.html |CUMSUM|> calculates the prefix sum. You can run this in parall...

fast 12 Jahre vor | 0

Beantwortet
A communication mismatch error was encountered: The other lab became idle during labReceive.
This error essentially tells you that the other worker (or "lab") got to the end of the |spmd| block before it sent any data. A ...

fast 12 Jahre vor | 0

Beantwortet
What is wrong with my arrayFun function ?
The error message is trying to indicate to you that the body of your |arrayfun| function cannot contain either method calls on M...

fast 12 Jahre vor | 0

Beantwortet
How can I assign *.mat table to GPU variable
You can send data to the GPU using the <http://www.mathworks.co.uk/help/distcomp/gpu-computing.html |gpuArray|> function, howeve...

fast 12 Jahre vor | 1

| akzeptiert

Beantwortet
Why is MATLAB running slow? 2013a vs 2011a
I tried your code on my machine in both R2011a and R2014a, and as posted it is indeed quite a bit slower in R2014a. You can get ...

fast 12 Jahre vor | 1

| akzeptiert

Beantwortet
How to run a job on a remote cluster?
Parallel Computing Toolbox relies on Java, so you should omit the |-nojvm| flag from your |matlab| command-line.

fast 12 Jahre vor | 0

Beantwortet
Using LaTeX interpreter + saveas png/epcs inside a parfor loop
Unfortunately this is a known problem - the bug report is here: <http://www.mathworks.co.uk/support/bugreports/253188> (I know t...

fast 12 Jahre vor | 0

| akzeptiert

Beantwortet
SPMD vs PARFOR and Memory Usage
Here's a version that I've reworked quite a bit to use codistributed arrays hopefully a little more effectively (it runs about t...

fast 12 Jahre vor | 3

| akzeptiert

Beantwortet
GPU memory accessed even when code runs on CPU only (See attached code)
I'm still investigating this problem, but as a workaround you can also use the 'like' syntax which avoids the problem: z = ...

fast 12 Jahre vor | 0

| akzeptiert

Beantwortet
Shared cluster profiles between R2013b and R2012a
You can create a cluster profile by creating the cluster object you wish to use manually from the command-line and then using <h...

fast 12 Jahre vor | 0

| akzeptiert