Thread Subject:

R2012b unresponsive: huge lag during debugging

Set "break on error". R2011b jumps immediately to the offending line. R2012b takes several seconds.

Stop debugging at that point: same story. R2012b takes anywhere from 10 to 30 seconds, while the 2011 release response immediately.

What happened in the 2012 release? Why is it so unresponsive?

Any ideas?

"Petr Krysl" wrote in message <k7eaf5$q7d$1@newscl01ah.mathworks.com>...
> Set "break on error". R2011b jumps immediately to the offending line. R2012b takes several seconds.
>
> Stop debugging at that point: same story. R2012b takes anywhere from 10 to 30 seconds, while the 2011 release response immediately.
>
> What happened in the 2012 release? Why is it so unresponsive?
>
> Any ideas?

Huge overload of Java shit?

"Budias Aao" wrote in message <k7ebgv$1ic$1@newscl01ah.mathworks.com>...
> Huge overload of Java shit?

Unfortunately, that is a possibility. What does the MathWorks crew have to say to this?

"Petr Krysl" <pkryslNOSP@Mucsd.edu> wrote in message
news:k7eaf5$q7d$1@newscl01ah.mathworks.com...
> Set "break on error". R2011b jumps immediately to the offending line.
> R2012b takes several seconds.
> Stop debugging at that point: same story. R2012b takes anywhere from 10 to
> 30 seconds, while the 2011 release response immediately.
> What happened in the 2012 release? Why is it so unresponsive?
>
> Any ideas?

This is probably going to be difficult to diagnose over the newsgroup.
Please contact Technical Support and work with them to determine the cause
of this slowdown.

--
Steve Lord
slord@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

"Petr Krysl" wrote in message <k7f3c6$qbd$1@newscl01ah.mathworks.com>...
> "Budias Aao" wrote in message <k7ebgv$1ic$1@newscl01ah.mathworks.com>...
> > Huge overload of Java shit?
>
> Unfortunately, that is a possibility. What does the MathWorks crew have to say to this?


I believe that this is an artifact of the new toolstrip. I've reported this as far back as the new desktop beta in Feb 8 (way before R2012b). I reported it again in the R2012b pre-release (1-IZF2W8), back in July. I got a response from MathWorks that this was indeed traced to a problem in the new Desktop, and that it is expected to be solved in the upcoming official release of R2012b. It has indeed improved in the official R2012b release, but apparently not entirely resolved.

Yair Altman
http://UndocumentedMatlab.com

"Petr Krysl" wrote in message <k7eaf5$q7d$1@newscl01ah.mathworks.com>...
> Set "break on error". R2011b jumps immediately to the offending line. R2012b takes several seconds.
>
> Stop debugging at that point: same story. R2012b takes anywhere from 10 to 30 seconds, while the 2011 release response immediately.
>
> What happened in the 2012 release? Why is it so unresponsive?
>
> Any ideas?

Hi Petr,

I would recommend bumping up your Java heap Space a fair amount. I have found this desluggishifies 12b on my systems.

"Sean de Wolski" <sean.dewolski@mathworks.com> wrote in message <k7gvv8$sm3$1@newscl01ah.mathworks.com>...
> "Petr Krysl" wrote in message <k7eaf5$q7d$1@newscl01ah.mathworks.com>...
> > Set "break on error". R2011b jumps immediately to the offending line. R2012b takes several seconds.
> >
> > Stop debugging at that point: same story. R2012b takes anywhere from 10 to 30 seconds, while the 2011 release response immediately.
> >
> > What happened in the 2012 release? Why is it so unresponsive?
> >
> > Any ideas?
>
> Hi Petr,
>
> I would recommend bumping up your Java heap Space a fair amount. I have found this desluggishifies 12b on my systems.


Sean - the problem apparently stems from a Java memory leak. So while increasing the Java heap space provides temporary relief, it does not help when you have Matlab running for more than a few hours of intense debugging. At some point in time you'd reach the heap-space boundary. In addition, Matlab's process memory keeps growing, thereby slowing all other processes on your computer that have to contend for the same RAM. So I believe that this is not a valid long-term solution.

- Yair

Hi all,

We established in a separate email that Petr is indeed using the R2012b pre-release, where this was a known issue. Here is the snippet from his "ver" command --

MATLAB Version 8.0 (R2012b Prerelease)

So we do not believe the issue exists in the R2012b final release. Please let us know otherwise.

Thanks,
Jared

"Jared " <jared.macdonald.nospam@mathworks.com> wrote in message <k8dnqn$o9j$1@newscl01ah.mathworks.com>...
> Hi all,
>
> We established in a separate email that Petr is indeed using the R2012b pre-release, where this was a known issue. Here is the snippet from his "ver" command --
>
> MATLAB Version 8.0 (R2012b Prerelease)
>
> So we do not believe the issue exists in the R2012b final release. Please let us know otherwise.
>
> Thanks,
> Jared


Hi Jared,

Thanks for the followup. This may indeed be the case with Petr, but I'm afraid that I cannot fully confirm. In fact, on my *production* R2012b I still see noticeable hangs when I use the debugger and hit breakpoints. It is hard to compare response times, but my subjective impression is that it is significant compared to R2012a.

So sorry, but I'm not sure you've fully fixed the problem. Or maybe you've only fixed the major issue and not some other problem. Or maybe I'm just getting paranoid and seeing extra lags where there aren't any...

- Yair

"Yair Altman" wrote in message <k8eavj$97s$1@newscl01ah.mathworks.com>...
> In fact, on my *production* R2012b I still see noticeable hangs when I use the debugger and hit breakpoints. It is hard to compare response times, but my subjective impression is that it is significant compared to R2012a.

I also see the same behaviour. I am running the production version of 2012b, and I see very long lags, especially when I quit from a debugging session. When I first start Matlab, the lag is not so noticable, but after an hour or so of work, I start to see lags of 3-5 seconds as I enter or quit from the debugger.

The "ver" command tells me:
MATLAB Version: 8.0.0.783 (R2012b)

Regards,
   Dave

I am getting similar behavior, and I am wondering if this is the same thing you folks are experiencing?

MATLAB Version: 8.0.0.783 (R2012b)
OS: Windows 8 Pro

The Matlab IDE becomes unresponsive for several minutes. It is weird, because I can still select text windows, and I can execute commands along the top menu strip. However, I cannot type anything into any of the windows, the cursor doesn't show up, and no response to any keyboard commands. However, I can resume debugging through the

Seems to happen when I exit debug mode with 'Quit Debugging' or when i save/overwrite the executing .m file and it exits debug mode.

-Corey

"Petr Krysl" wrote in message <k7eaf5$q7d$1@newscl01ah.mathworks.com>...
> Set "break on error". R2011b jumps immediately to the offending line. R2012b takes several seconds.
>
> Stop debugging at that point: same story. R2012b takes anywhere from 10 to 30 seconds, while the 2011 release response immediately.
>
> What happened in the 2012 release? Why is it so unresponsive?
>
> Any ideas?

Hi there,

I had a similar problem, my MATLAB Version: 8.0.0.783 (R2012b) installed on Win-8 64bit core-i3 4GB laptop, became rather completely unresponsive without any error or break. I called MATLAB from python win32com.client.

The script worked fine on R2007a. and thanks to this topic I realized that by actually commenting out the plot() option, my script was again functional on R2012b.

So what I realize is that the developers should think about graphics.

hi everyone!

i was wondering if anything happened since the last posts. in fact i'm experiencing the exactly same issue as Dave and Yair (huge lags in respect to debugging and breakpoints or when stopping debug mode.. after working with matlab for a certain amount of time) and i'm also using the release/"production" version of 2012b..

does the issue still exist in 2013a? .. although i would prefer a fix instead of buying a new version! ;)

bye the way, might this issue be adressed in this bug report? -> http://www.mathworks.de/support/bugreports/details.html?rp=819115

"D" wrote in message <kltjeh$j89$1@newscl01ah.mathworks.com>...
> hi everyone!
>
> i was wondering if anything happened since the last posts. in fact i'm experiencing the exactly same issue as Dave and Yair (huge lags in respect to debugging and breakpoints or when stopping debug mode.. after working with matlab for a certain amount of time) and i'm also using the release/"production" version of 2012b..
>
> does the issue still exist in 2013a? .. although i would prefer a fix instead of buying a new version! ;)
>
> bye the way, might this issue be adressed in this bug report? -> http://www.mathworks.de/support/bugreports/details.html?rp=819115

It's written there: [ This bug was fixed as of R2013a(8.1). ]

Bruno

"D" wrote in message <kltjeh$j89$1@newscl01ah.mathworks.com>...
> hi everyone!
>
> i was wondering if anything happened since the last posts. in fact i'm experiencing the exactly same issue as Dave and Yair (huge lags in respect to debugging and breakpoints or when stopping debug mode.. after working with matlab for a certain amount of time) and i'm also using the release/"production" version of 2012b..
>
> does the issue still exist in 2013a? .. although i would prefer a fix instead of buying a new version! ;)
>
> bye the way, might this issue be adressed in this bug report? -> http://www.mathworks.de/support/bugreports/details.html?rp=819115


As far as I can tell, there is still a noticeable lag in R2013a, but I can't tell whether it is less than on 12b. It is certainly larger than on 12a and earlier releases.

This is a Matlab bug and MathWorks should fix it. It was reported over a year ago and I don't understand why it is not yet [fully] fixed.

I must also say that I find the lack of official MathWorks response in this thread inappropriate. Guys - if this is a bug then come outright and say it, provide information, and tell us what you plan to do about it - don't let us all hang in the air. This is not a minor bug or inconvenience; it is a major pain. It's your interest to sell license upgrade, not to force users to rely on older Matlab releases. Why do I have to spell this out?

Yair Altman
http://UndocumentedMatlab.com
 

"Yair Altman" <altmanyDEL@gmailDEL.comDEL> wrote in message
news:kltkg9$lvt$1@newscl01ah.mathworks.com...
> "D" wrote in message <kltjeh$j89$1@newscl01ah.mathworks.com>...
>> hi everyone!
>>
>> i was wondering if anything happened since the last posts. in fact i'm
>> experiencing the exactly same issue as Dave and Yair (huge lags in
>> respect to debugging and breakpoints or when stopping debug mode.. after
>> working with matlab for a certain amount of time) and i'm also using the
>> release/"production" version of 2012b..
>>
>> does the issue still exist in 2013a? .. although i would prefer a fix
>> instead of buying a new version! ;)
>>
>> bye the way, might this issue be adressed in this bug report? ->
>> http://www.mathworks.de/support/bugreports/details.html?rp=819115
>
>
> As far as I can tell, there is still a noticeable lag in R2013a, but I
> can't tell whether it is less than on 12b. It is certainly larger than on
> 12a and earlier releases.
> This is a Matlab bug and MathWorks should fix it. It was reported over a
> year ago and I don't understand why it is not yet [fully] fixed.

Have you reported this both here and to Technical Support? Reporting it on
CSSM is sufficient to inform other users; to ensure it gets on our
developers' plate, reporting it to Technical Support is necessary.

> I must also say that I find the lack of official MathWorks response in
> this thread inappropriate. Guys - if this is a bug then come outright and
> say it, provide information, and tell us what you plan to do about it -
> don't let us all hang in the air. This is not a minor bug or
> inconvenience; it is a major pain. It's your interest to sell license
> upgrade, not to force users to rely on older Matlab releases. Why do I
> have to spell this out?

The official avenue to obtain technical support from MathWorks is to contact
our official Technical Support group. My signature includes instructions on
how to do so. If you check back in this thread to message 4 I recommended
that people who are experiencing this contact Technical Support the day
after the thread started.

Determining the root cause of an issue this general and developing a
solution to that root cause is going to be extremely difficult unless the
right people are involved. I don't work on the Desktop team, and I'm not
sure how many (if any) of the Desktop team read CSSM regularly. If you post
here the developers _may_ see it or may not. If you contact Technical
Support the Support staff can work with you and the appropriate developers
to figure out what's going on and the right way to fix it.

The cause of this issue in MATLAB. It may be a general Java issue. It may be
an interaction between MATLAB, Java, the operating system, other pieces of
software, etc. that we can't resolve on our own. I don't know. The Support
staff may know, or they can ask the developers who should know (or will be
able to find out.)

--
Steve Lord
slord@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com