A few days ago I wanted to upgrade Julia and all dependencies to 1.0. Therefore I deinstalled all old versions and installed the JuliaPro distribution 1.0.2.1. So far everything went good, Juno ran without problems and I had no complains whatsoever. Yesterday however out of the blue sky, Atom was extremely laggy. I mean by that, that whenever something is executed Atom freezes or reacts only several seconds after the expression is already evalued. Sometimes even this happens:
The only thing I changed on my computer before the lags appeared was a Java update.
I wondered whether this could have been due to having much data saved in variables but neither was my RAM to full (<80%), nor was the Julia language process using much RAM space. Additionally, in the console my file runs flawlessly. So I tried rebooting the computer several times and not using some Julia packages, as well as Atom packages, but it did not change anything.
After that I supposed it could be due to some problem of the JuliaPro distribution as a whole and therefore reinstalled Atom + uber-juno + Julia manually. Again, this didn’t help.
It may be useful to know that at the beginning of the session I sometimes get this error: julia client internal error TypeError: in FixedPoint, in T, expected T<:Integer, got Type{Char}
I would really appreciate it if somebody would have an idea about what is going on…
I think I figured it out. One reason was indeed the opened workspace pane. But in fact closing the workspace pane only lowers the intensity of lags. The other cause is that the results also get displayed behind the code which is of course effectively the same problem. Therefore by avoiding that too much code gets displayed the lags disappear.
Hm, do you only get lags when getting certain types of results back or always when you have a large number of simple results? E.g. if you have a screen full of 1+1 and execute all of them sequentially, does Atom lag?
Would be awesome if you could give me some more details or even a short screencapture to help me figure out what’s going on.
To clarify matters: What I’m currently programming is a framework for solving and optimizating certain nonlinear equation systems for macroeconomic models. To ensure compability with other languages I take a text file as input and then parse it so that I get Julia functions of it. And some of those functions can get relatively big, around 1000 lines of code. For debugging purposes I often wanted to see the output of the parsing process before I executed eval(). And keeping that code in the workspace pane and/or displaying it as result in the editor causes extreme lags. What seemed strange to me is that, as I already stated above, the problems didn’t arise due to the pure amount of consumed storage. Instead, it seems to be a Juno specific problem.
I have similar problems with Juno, it is very anoying (just compared to MATLAB).
For example for debugging you should enter Juno.@enter functionName(args). And then in debugging it is extremely slow and sluggish.
