Matlab died when randomforest finished its process

[GoogleCodeExporter]

What steps will reproduce the problem?
 I am using the randomforest on 64bit Linux machine with Matlab version
7.9.0.529 (R2009b) through SSH Secure Shell Version 3.2.9. 
Everything is fine but after randomforest finished its job, Matlabe is just
stuck. This means that Matlab does not respond. 
I looked at the process, but it seems Matlab is just stuck but the process
is still alive, but I cannot do anything else, so I have to disconnect the
terminal and log in again. 
Once I disconnected the terminal then Matlab process is killed and there is
no dump file. 

Original issue reported on code.google.com by korjcje...@gmail.com on 21 Apr 2010 at 9:09

[GoogleCodeExporter]

can you give the part of the data and code for which this happens?

also did you run randomforest before in the same matlab instance and cancel it 
by any chance 
via ctrl+C?

thanks

Original comment by abhirana on 21 Apr 2010 at 9:13

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

I ran below codes, but it happens again.

>> train_data = rand(10,100);
>> train_label = [ones(5,1); ones(5,1)*-1];
>> Model = classRF_train(train_data, train_label, 0, 0);

Yes, I think I canceled job with ctrl+C.
However, I logged out and ran it again, but still it happens.

Thanks. 

Original comment by korjcje...@gmail.com on 23 Apr 2010 at 10:01

[GoogleCodeExporter]

well i am not able to reproduce the problem on linux with ssh(ing) from the 
same machine. And 
also from windows to a linux machine via putty

i can suggest the following: try using putty as a ssh client or use `screen' 
command to spawn 
matlab and see.

btw is the X being forwarded on windows via SSH? try using without the 
forwarding

Original comment by abhirana on 23 Apr 2010 at 10:15

[GoogleCodeExporter]

I just found that this happens when I run matlab with nodisplay option.
In the command line.
>> matlab -nodisplay

I run the matlab on console mode then it happens. 

Thanks.

Original comment by korjcje...@gmail.com on 26 Apr 2010 at 3:53

[GoogleCodeExporter]

sorry i could not get to you earlier.

if it is same to you, i.e getting matlab on ssh, try with matlab -nojvm and 
with a different ssh 
client

i am still not able to reproduce the same error on my machine with and without 
ssh and using 
the nodisplay option

Original comment by abhirana on 7 May 2010 at 2:43

[GoogleCodeExporter]

I tried -nojvm with putty but the problems is still the same.

Thanks.

Original comment by korjcje...@gmail.com on 7 May 2010 at 5:02

[GoogleCodeExporter]

I too am getting exactly the same error as korjcjeong. I am using 7.5.0.342 
(R2007b).
In particular, the problem only happens for classification (i.e. when using 
classRF_train.m) and not for regression (i.e. refRF_train.m is OK).  I wonder 
if there is an issue using calloc [e.g. double* countts = (double*) 
calloc(nclass * nts,sizeof(double)); on line 189] instead of letting MATLAB 
doing the memory managment, as is recommended, or a memory leak.

Original comment by erik.and...@gmail.com on 30 Jul 2010 at 7:19

[GoogleCodeExporter]

so are you also using it in the same setting of ssh's and using matlab? i was 
never able to reproduce the same problem.

well matlab variables cannot be used in the C code if they are to be written 
to, they are good only for reading, thats why there are allocs in C.

Original comment by abhirana on 31 Jul 2010 at 9:07

[GoogleCodeExporter]

Hi,

I also have the same problem. Matlab freezes after it returns from 
classRF_train when I run matlab through ssh without GUI.
I tried this on SUSE and REDHAT and both get stuck after classRF_train and also 
classRF_predict.
It works without problem when I run matlab with GUI.

Thanks,
MJ

Original comment by m.se...@gmail.com on 28 Jun 2012 at 5:03

[GoogleCodeExporter]

@MJ

did you try out my earlier suggestion in the other thread?

possible issue might be that particular version of gfortran and g++ don't work 
well in conjunction with malloc and matlab. try to sync with the svn source 
which replaces mallocs by matlab based malloc mxMalloc 
http://code.google.com/p/randomforest-matlab/source/checkout

Original comment by abhirana on 28 Jun 2012 at 10:36

[GoogleCodeExporter]

I have solved this problem by disabling X11 tunneling.

Original comment by korjcje...@gmail.com on 29 Jun 2012 at 9:05

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

I forgot to mention that, yes I updated gfortran and g++ to 4.4.1 version and I 
use the latest version on the repository but still I have the problem. I'm 
trying to reproduce the problem to find out more about the issue. I'll provide 
more information if I find any.

Thanks,
MJ

Original comment by m.se...@gmail.com on 29 Jun 2012 at 10:32

[GoogleCodeExporter]

hi mj

what options do you use to run matlab in ssh?  korjc (thanks!) reported it 
might be due to X11 tunneling. the other option is to see if one or a 
combination of the startup options is causing issue (comment-5), startup 
options include either -nojvm -nodisplay

if you a test code i can try it on my machines and try to see if i can 
reproduce them on debian.

Original comment by abhirana on 30 Jun 2012 at 1:16