R2015a super slow on Linux

I can't figure out what's wrong. My machine isn't that slow. It's a i5-2430M @ 2.4ghz x4, 6g ram and a 500gb ssd. I had R2014a running fine. I could run 45 minute simulations w/o problems and even make use of the parallel pool. I installed R2015a and I constantly get java heap errors and the processor is being stressed out constantly. Often it will freeze to a near halt where commands will register 4-5s after issuing them and writing scripts will take forever. I've tried allocating more memory to the heap but I can't see why the majority of my programs would need that much. I just now started it up and received an error w/o doing anything - '...GC overhead limit exceeded'. If anyone has any ideas or if additional information is needed please let me know. It's very painful to resort to ssh'ing into my school's machines or having to do all my work at a lab.

[Walter Roberson]

One facet that comes to mind is that R2015a would use Handle Graphics 2 but R2014a would use the original Handle Graphics.

I am seeing hints in posting that OpenGL might also have had some changes as of R2014b; have you tried switching to opengl hardware mode? You would need a startup flag for that on Linux; see here

[Lian]

I have the same question. I just have a linux machine with intel Xeon 2.4Ghz 6 core processor and 2015a runs the same script quite more slowly than my mac pro with 2.4Ghz intel Core i7 and 2011b. The OpenGL issue does not seem to fix the problem.

[Patrick]

Yeah, OpenGL didn't seem to resolve it for me.

[Patrick]

so... Anyone else? This is sort of important.

[Patrick]

I still have this problem...

[Walter Roberson]

A bit of a stretch, but is it possible that you are running against power supply issues? Hypothetically R2015a could be enough more efficient to be stressing the system. It doesn't seem especially likely, but just maybe....

[Ahmed Sabir]

I'm having the same problem, my spec (SSD 500G, i7 @ 3.4ghz x8, 23g Ram). I can't figure out why matlab is super slow and not optimize enough in Linux.

[Jan]

@Ahmed: Do you see that this thread does not contain any specific detail yet? It matters which commands are slow and which Matlab version you are running. Perhaps you have written an infinite loop or forgot a pre-allocation. It is impossible to suggest an improbvement based on the information: "Matlab is slow". The forum cannot suggest to use the "-muchfaster" flag in the command line, unfortunately.

Please open a new thread and post the relevant details.