Silent crash when trying to set large numpy array

When I’m trying to set some large numpy array variables in Matlab through the set_variable method, my Matlab server close and the notebook stays idling forever. Here is a minimal code for reproducing this error (I’m running it in a notebook):

import numpy as np
from pymatbridge import Matlab
import sys
mlab = Matlab(maxtime=60) #Matlab(matlab='C:/Program Files/MATLAB/R2015a/bin/matlab.exe')
mlab.start()

print("Before")
sys.stdout.flush()
mlab.set_variable("in0", np.zeros((3814400,1)))
print("After")
sys.stdout.flush()

This gives me the output:

Starting MATLAB on ZMQ socket tcp://127.0.0.1:51192
Send 'exit' command to kill the server
............MATLAB started and connected!
Before

And at that point, the Matlab server window close automatically, “After” never get printed and the notebook stays idling (i.e. it shows “In [*]” forever) forcing me to reboot the kernel.

[alexvorndran]

I ran into the same problem while passing around some larger numpy arrays. I’m working with a 64bit version of Matlab R2013a on Linux Mint 14.1 (IcedTea Java - I don’t know the precise version). To some extent you can stop this from happening if you increase the Java Heap Memory, that Matlab is allowed to use.

For further information on this topic and an instruction, see the Matlab documentation here and an older Mathworks blog post here.

[arokem]

I’m sorry: I am not really using Matlab much these days, and have neither the bandwidth, nor the strong motivation to dig into these issues. I suspect that it might have to do with hard-coded limitations (maybe this? https://github.com/arokem/python-matlab-bridge/blob/master/pymatbridge/messenger/src/messenger.c#L8)