Recently, I was faced with a new error in MATLAB. It says "Out of memory." I think there should be a way to increase it's memory. Any suggestions?
I have the following suggestions for you:
You can see my answer to a similar question: "For Matlab program, how I can solve the problem of "out of memory"? " on research gate (same link Harald recommends. In that answer I focused on MATLAB program structure to avoid this problem. The advices given by Harald are worthy to consider as well.
However, on the device management side, some tricks can help like:
1- Give higher priority to Matlab.exe task. You can do that by going to task manager, Processes tab, right click the Matlab.exe Task, select priority and set it to higher priority (say real time), this tells Windows to allocate more resources to this process.
2- Increase the page file size of your applications in general. You can do this by right clicking MyComputer ->properties->Advanced System Settings ->Advanced-> Performance->Virtual Memory (change..). Then the tick from the Automatic .... and set the initial and maximum page size to say 10000 MB.
3- While in 2 at Performance, make sure that the it is set for "Programs" and not for "Background services". "Programs" should be the default.
4- Also make sure while doing step 2 that the >Visual effects tab of "performance options" is set to "Adjust for best performance" . This will remove lots of animation, shades, fades and and other visual effects from your screen (the next time you re-boot) but free some memory for the applications including the MATLAB. I used this trick with my older laptop which was 32 bits OS and used to have a profound impact on applications performance.
After all don't forget to give us the feedback what worked for you from either my recommendations or those of other colleagues for the benefit of all. Thank you.
Here's a link showing how exactly it works: http://www.mathworks.in/help/distcomp/matlabpool.html But the gist is when you simply type matlabpool open in the command window of matlab and write "parfor" in place of "for" in the loops, it makes all the 4 cores of your system work in parallel for producing matlab output. This will reduce load and improve speed...
It should be noted that the "matlabpool", and "parfor" commands are part of the Parrallec Computing Toolbox and will not work unless you have it installed.
I have the following suggestions for you:
You can see my answer to a similar question: "For Matlab program, how I can solve the problem of "out of memory"? " on research gate (same link Harald recommends. In that answer I focused on MATLAB program structure to avoid this problem. The advices given by Harald are worthy to consider as well.
However, on the device management side, some tricks can help like:
1- Give higher priority to Matlab.exe task. You can do that by going to task manager, Processes tab, right click the Matlab.exe Task, select priority and set it to higher priority (say real time), this tells Windows to allocate more resources to this process.
2- Increase the page file size of your applications in general. You can do this by right clicking MyComputer ->properties->Advanced System Settings ->Advanced-> Performance->Virtual Memory (change..). Then the tick from the Automatic .... and set the initial and maximum page size to say 10000 MB.
3- While in 2 at Performance, make sure that the it is set for "Programs" and not for "Background services". "Programs" should be the default.
4- Also make sure while doing step 2 that the >Visual effects tab of "performance options" is set to "Adjust for best performance" . This will remove lots of animation, shades, fades and and other visual effects from your screen (the next time you re-boot) but free some memory for the applications including the MATLAB. I used this trick with my older laptop which was 32 bits OS and used to have a profound impact on applications performance.
After all don't forget to give us the feedback what worked for you from either my recommendations or those of other colleagues for the benefit of all. Thank you.
The best way is to preallocate the memory if you know the size of matrices that you working. This create a contiguous block to store the data. Other thing, is disable java (which is a memory hogger) like is and this works better if you are only analyzing the data not using visualizations.
Can ,any one solve my Problem?
I am using MATLAB R2010a,32-bit....for my work. When my model is running, it's going to stop and indicates "Memory Allocation ERROR"..How can I over come this problem and my model will run properly.
Some useful suggestions:
http://www.mathworks.com/help/matlab/matlab_prog/resolving-out-of-memory-errors.html
none of the above suggestions resolved my problem. Can any one help me out to solve this problem
There are a couple of things you can try. First clear whatever variable you don't need anymore in your simulation with "clear" command. I tend to even delete counters immediately after use. Use "pack" command to compress variables in memory and make more free space. You can also use "zeros" function to predetermine huge the variables/matrixes. There is also an internal memory in MATLAB which can be increase, but I don't think that's the bottle neck in your simulation since if that was the case MATLAB would close itself without showing any error. In addition, you need to consider the fact that depending on the MTALAB and OS version there is a limitation on accessible memory for MATLAB. The maximum accessible memory is attainable while both MATLAB and OS are 64-bit on a 64-bit CPU. Otherwise, the accessible memory is limited to ~4 GB regardless of the hardware you've got (http://au.mathworks.com/help/matlab/matlab_prog/resolving-out-of-memory-errors.html).
Apart from these things, you can profile MATLAB memory usage to get a clear understanding on which part of your code requires optimization the most. Here is an example: http://undocumentedmatlab.com/blog/internal-matlab-memory-optimizations
Hope this helps!
Just remembered something, when dealing with big matrices, try to split the algorithm to several parts in a way that if the out-of-memory error occurred the commutation can be continued on another MATLAB session. You may use "save" command to pass all the variable to the new MATLAB session. This is what I do when I optimise my code but still need more memory.
i also have this problem when load a big file in matlab , i don't know what shall i do >>>any suggestions?
Before the network training....use this line...
net.efficiency.memoryReduction = n;
increasing n-->10-->20-->40----> 100 use less memory but it will slowdown your computation... but after-all in the end you can find the answer...
Use proper value of n...
Actually I dont know how to increase the memory, but, If you use 64 bit MATLAB with 64 bit operating system then you can have 8TB of memory, check the link: http://www.mathworks.com/help/matlab/matlab_prog/resolving-out-of-memory-errors.html
If you can't increase the memory size or above methods don't work, I would suggest you to optimize your code. For example, if you have to add two large sparse matrices, partition each matrix into sub-matrices and perform the addition separately. Later, assemble the resultant sub-matrices. Please remove the unwanted variables in the process from the workspace.
- Badges
- Science topic