General matlab hints: http://math.aalto.fi/opetus/Mattie/MattieO/matlab.html
Installation and license activation on staff-owned computers¶
Matlab academic license permits installation on home computers for university personnel. Triton MDCS workers are available to anyone with a Triton account, which means the workers can be utilized from personal laptops as well.
Log into http://download.aalto.fi/ with your Aalto account. Look for the link Software for employees’ home computers which will take you to the Matlab download links. Download the UNIX version for Linux and OSX or the separate separate image for Windows.
The ISO image can be burned on a DVD or mounted on a virtual DVD drive.
Select the installer options as shown in the screenshots.
Mathworks account is required to continue with the installation.
- Enter your account information in the installer to log in. If the password has been lost, Click on the Forgot your password? option to receive your password in email. OR
- Register to Mathworks with the installer.
- Click on I need to create an account.
- Enter your name and email address. To be recognized as Aalto academic user the email address must end in one of aalto.fi, tkk.fi, hut.fi, hse.fi, hkkk.fi or uiah.fi domains.
- The installer will ask for an activation key, which is shown here in the last screenshot.
You may leave out unnecessary toolboxes and change the installation location. Remember however, that the Parallel Computing Toolbox is necessary to run any Matlab batch jobs on Triton.
Matlab freezes with Out of Memory errors¶
Q: Matlabs freezes and I get errors like this. What to do?:
Exception in thread "Explorer NavigationContext request queue" java.lang.OutOfMemoryError: GC overhead limit exceeded at com.mathworks.matlab.api.explorer.FileLocation.<init>(FileLocation.java:89) at com.mathworks.matlab.api.explorer.FileLocation.getParent(FileLocation.java:126) ... ... ...
A1: Add more memory in Home -> Preferences -> General -> Java Heap memory
A2: Can you free up memory in your code sooner using the
Q: is there functional GPU acceleration? Does the acceleration even work?
A: run code:
>> g = gpuDevice; >> ng
A2: Just query some feature:
>> fprintf('%s\n', g.ComputeCapability)
a3: Show multiple devices if found:
>> for ii = 1:gpuDeviceCount g = gpuDevice(ii); fprintf(1,'Device %i has ComputeCapability %s \n', ... g.Index,g.ComputeCapability) end