Answered
speed up computation of huge matrix?
As Jan points out, this is very large data. One option is to use <http://www.mathworks.com/help/toolbox/distcomp/bquvksw.html di...

14 years ago | 2

Answered
Can't pass COM object handle into spmd body
As you've discovered, there are some limitations with passing handle-type objects into the body of SPMD blocks. This is because ...

14 years ago | 0

Answered
Parfor loop worker utilization problems
How many iterations are your PARFOR loops running over? If the number is small, then the work may end up not being distributed c...

14 years ago | 0

Answered
Loop through files of remote server
Perhaps a PARFOR loop on the cluster would be simplest. Something like this: function x = doStuff % list the files on...

14 years ago | 1

| accepted

Answered
parallellisation of a large array
GPUs are great at elementwise computations, by putting your data into a <http://www.mathworks.com/help/toolbox/distcomp/bsic4fr-...

14 years ago | 1

Answered
MDCS Homogeneous cluster configuration
It's really about the workers - there are some constraints from the MPI library we use (MPICH2) about the operating systems and ...

14 years ago | 0

Answered
GPU and Built-In Functions
What you're seeing is a single thread of execution on the CPU working hard to keep the GPU busy. This is expected behaviour.

14 years ago | 1

| accepted

Answered
Processing blocks of data in parallel
Your best bet is to try and restructure your code to use a PARFOR loop. For example, you could do something like this: N = ...

14 years ago | 1

| accepted

Answered
Parallel execution of external application with different parameters
I would use the TEMPNAME function, as recent versions of MATLAB generate names that are highly unlikely to clash.

14 years ago | 0

Answered
i want to run parfor loop.... but there is a problem like slice... help me.....
You're trying to write the results of your PARFOR loop into 'B'. The rules for doing this mean that you can only index 'B' with ...

14 years ago | 2

Answered
parallel.gpu.CUDAKernel memory types
Currently it's not possible to use either texture or constant memory with CUDAKernel objects. The memory caches in recent NVIDIA...

14 years ago | 0

Answered
Matlab pool error
It sounds like your 'DataLocation' property on the 'local' scheduler has got an invalid value in it ('4') - this needs to be a d...

14 years ago | 1

Answered
memory allocated with mxMalloc
Do you mean a 'batch job' created using the BATCH function in Parallel Computing Toolbox? If so, then no, the workers executing ...

14 years ago | 0

| accepted

Answered
Matlab 2011b and Cuda
As Walter says, there are some relatively new CUDA drivers available for Linux, it's definitely worth getting the latest. If you...

14 years ago | 0

Answered
parfor-loop: How can I increase the number of workers?
You should be able to increase the number of workers for the 'local' scheduler, but the maximum is 8 (for R2011a) or 12 (for R20...

14 years ago | 0

Answered
old batch job blocking new ones
When you open a MATLABPOOL, you are running a job on your scheduler. That job is connected up to your MATLAB client to provide w...

15 years ago | 0

| accepted

Answered
Error using UseParallel and fmincon
GLOBAL (and PERSISTENT) variables are not shared between your MATLAB client session and the workers executing the body of your ...

15 years ago | 0

Answered
Why do I receive this error? Error using ==> psoptimset at 297 Unrecognized parameter name 'UseParallel'.
Parallel Computing Toolbox support was added to Global Optimization Toolbox starting in R2008a, as per the release notes here: ...

15 years ago | 1

Answered
Distributing Arrays with parfor
Distributed arrays are designed to be used with SPMD blocks, not PARFOR loops. In general, it is a mistake to use a distributed ...

15 years ago | 1

Answered
Doubts with parfor loop
The problem is that you're using the whole value of 'baseDatosAct' in the line: veDelanteMiaTemp=vehDelanteMia(baseDatosA...

15 years ago | 0

Answered
matlab seg fault on distributed computing cluster
Parallel Computing Toolbox is only designed to work with the same release of MATLAB Distributed Computing Server, as per the <ht...

15 years ago | 0

Answered
access to a gpuArray from a MEX file?
You're right that you need Parallel Computing Toolbox to access gpuArray functionality. We don't offer a MEX interface with gpuA...

15 years ago | 0

| accepted

Answered
parfor worker affinities
PARFOR is designed for order-independent loop iterations, and there's no way to control the placement of the iterates. If the "c...

15 years ago | 2

Answered
Problem in running multiple parallel jobs on cluster
Are you including the "matlabpool" command in your code to run on the cluster? If so, that's not the right way, you should be su...

15 years ago | 1

| accepted

Answered
Issues using PARFOR with large object array (slow execution time)
As you increase the number of workers, the amount of communication doesn't increase (for sliced inputs/outputs), but the amount ...

15 years ago | 1

| accepted

Answered
Run function on GPU, Matlab 2011a
Yes, you can do that. The GPU variant of arrayfun has been slightly expanded from the standard MATLAB version to allow scalar ex...

15 years ago | 2

Answered
Parallel speedup problem with spmd
There is definitely some overhead to entering and leaving an SPMD block. On my machine with "matlabpool local 2", it's about 0.0...

15 years ago | 0

Answered
How do you tell local workers to use the session of Simulink which is already open?
MATLAB workers are completely separate processes, and as such, cannot share Simulink data. One option you have is to use MAT...

15 years ago | 0

| accepted

Answered
Uploading small vectors to GPU is faster than directly creating them
The parallel.gpu.GPUArray.* build functions in R2010b were provided mostly to allow you to avoid host allocation prior to copyin...

15 years ago | 1

| accepted

Answered
for loop in parfor
Yes, it's possible. For example, parfor ii=1:10 for jj=1:10 x(ii,jj) = rand end end Note that th...

15 years ago | 1

| accepted