bad_alloc error

18 views
Skip to first unread message

l.main...@gmail.com

unread,
May 3, 2013, 11:20:54 AM5/3/13
to ukb...@googlegroups.com, Daniele Vannella
Hello,
I'm an IT student and I am working on my thesis with Professor Navigli at Sapienza University in Rome.
I'm trying to compile a very large graph (about 90 millions of edges) but i get the following error: 

std::bad_alloc in line 19897599

Of course, that line does not contains any error and the whole source file seems to be ok.

I've also tried to run ukb in a pc provided with 64 Gb of RAM, with the following result:

std::bad_alloc in line 39856706

So it seems to be a problem due to a lack of memory, but the source file size is just 4,7 Gb. Yet, if I erase some rows, the returned line number changes, and to be more precisely it only decreases, but there is no way to avoid the error is thrown.

So I'd like to know if you have any suggestions.

Thank you in advance.

Regards

Leonardo Mainardi

Aitor Soroa

unread,
May 3, 2013, 2:17:15 PM5/3/13
to ukb...@googlegroups.com, Daniele Vannella
Hi Leonardo,

I guess that you have reached the maximum memory limit. If ukb is
compiled on a 32 bit system it can only manage up to 4Gb RAM (as any 32
bit application). To overcome this problem you'll have to compile (an
run) ukb on a 64 bit system.

Also, I strongly recommend you to use ukb 2.0 for dealing with such
large graphs (version 2.0 uses much less memory for storing the graphs).
Maybe you are already using version 2.0, but I thought it is worth
mentioning.

best,
aitor
> --
> You received this message because you are subscribed to the Google Groups "ukblist" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ukblist+u...@googlegroups.com.
> To post to this group, send email to ukb...@googlegroups.com.
> Visit this group at http://groups.google.com/group/ukblist?hl=en.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
ondo izan
aitor

l.main...@gmail.com

unread,
May 5, 2013, 9:16:13 AM5/5/13
to ukb...@googlegroups.com, Daniele Vannella, l.main...@gmail.com
Hi Aitor,
thank you for your help. I've tried to follow your suggestion, but i haven't succeeded in compiling ukb, because of some problems due to boost libraries.
I have installed boost 1.53 (the latest version) but ukb doesn't "like" this version.
So I have installed some previous versions (1.43, 1.44, 1.49) but when i execute 'make install', i always get the errors you can read in the attached files.
Any ideas?
Or, can i get a precompiled 64-bit version of ukb from somewhere?

Thank for your attention.

Best,
Leonardo Mainardi
make install (using boost-1.43)
make install (using boost-1.49)

Aitor Soroa

unread,
May 6, 2013, 3:02:21 AM5/6/13
to ukb...@googlegroups.com, Daniele Vannella, l.main...@gmail.com
Hi Leonardo,

could you try boost 1.46 ? This is the version of boost I've tried
myself, and it does compile.

I'm trying to understand the error message you send me, but still don't
know what's happening. Btw, which version of g++ are you using ?

best,
aitor

l.main...@gmail.com

unread,
May 6, 2013, 10:59:34 AM5/6/13
to ukb...@googlegroups.com, Daniele Vannella, l.main...@gmail.com
I've just tried to install boost 1.46, but i get the same errors. My g++ version is 4.7.2. I've also tried to install boost 1.46 via apt, but with no success. Do you by any chance have a precompiled 64-bit version of ukb you can send to me?

Thank you again,
Leonardo 

On Friday, May 3, 2013 5:20:54 PM UTC+2, l.main...@gmail.com wrote:

Aitor Soroa

unread,
May 7, 2013, 12:49:22 PM5/7/13
to ukb...@googlegroups.com, Daniele Vannella, l.main...@gmail.com
Hi Leonardo,

I've precompiled a 64 bit version. You can download it from here:

https://siuc05.si.ehu.es/~ccpsoeta/ukb_static64.tar.bz2

best,
aitor

l.main...@gmail.com

unread,
May 7, 2013, 7:56:14 PM5/7/13
to ukb...@googlegroups.com, Daniele Vannella, l.main...@gmail.com
It works! Great!
I don't know how to thank you enough for all your help.

Best,
Leonardo

On Friday, May 3, 2013 5:20:54 PM UTC+2, l.main...@gmail.com wrote:
Reply all
Reply to author
Forward
0 new messages