Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-discussion] Undelivered Mail Returned to Sender

Tania Tudorache tudorache at stanford.edu
Fri Sep 8 10:24:24 PDT 2006


My computer configuration is similar to Raj's (Intel Pentium M, 2.2GHz, 
2GB RAM) . With 700MB of heap size I could load the 300.000 classes 
ontology in memory.

Anyway, the idea is that at some point, you will reach the limit of the 
heap space that Java can allocate (I think it is 1.6 GB on Windows), so 
you won't be able to load into memory a huge ontology. That is why you 
have to use the OWL db backend.

Of course, everything depends on the scenario that you have. If you 
start to develop a new ontology, and you know that it will be huge, then 
you should start with a project using the OWL db backend. In this case, 
you don't have any limitation on the size of the ontology.

If you already have a big OWL ontology as a file and you want to import 
it into an OWL db, so that you can use it with the OWL db backend, there 
are some solutions:
1. If you don't have enough memory, you can ask somebody (with enough 
memory) to load the owl file in Protege and convert it into a db project 
(this is a one time job). And he/she can get you a db dump that you can 
then use.
2. Ideally, you could use a converter that takes an OWL file and 
converts it directly to an OWL db without loading it completely in 
memory. We are working on this, but I cannot give you an exact date when 
it will be released.

Tania


Mudunuri, Raj wrote:

>Hi Congmin,
>
>I have 2G RAM, where I have allocated 1G of heap space for Protege... It took around 100 secs to load all the triples, then for around 50 secs it did nothing (looked dead) and then the UI came into normal stage... so, totally it took around 150 secs in my system to load your ontology with 300,000 concepts... mine is an Intel Pentium D, with 2.8 GHz and 2 GB RAM...
>
>Cheers,
>Raj
>
>
>
>-----Ursprüngliche Nachricht-----
>Von: protege-discussion-bounces at lists.stanford.edu [mailto:protege-discussion-bounces at lists.stanford.edu] Im Auftrag von Congmin min
>Gesendet: Freitag, 8. September 2006 04:43
>An: tudorache at stanford.edu; User support for Core Protege and the Protege-Frames editor
>Betreff: Re: [protege-discussion] Undelivered Mail Returned to Sender
>
>Tania:
>
>   Could you let me know how large is your CPU and physical MEMORY? I
>increased the heap size to 1GB, but had the same problem. Probably I should
>install more CPU and RAM. When I try loading, the backround message showed
>that everything has been loaded into memory compeletely, but it just didn't
>come out on the Protege UI. And then looks dead.
>
>Thanks,
>congmin
>
>On 9/7/06, Tania Tudorache <tudorache at stanford.edu> wrote:
>  
>
>>Congmin,
>>
>>I did receive your email. Indeed this is a problem, that you have to
>>first load the OWL file in memory and then to export it to an OWL
>>database. We are looking at ways on how to do a streaming writing to the
>>database directly.
>>
>>Until then, you need to increase the heap size, if you need to
>>accomodate very large ontologies.
>>
>>Tania
>>
>>Congmin min wrote:
>>
>>    
>>
>>>Hi, Tania:
>>>
>>>      I am attaching three files for your testing:
>>>      The .tar.gz contains three owl files I generated. Both are very
>>>simple owl files, only containing one class and a number of instances
>>>of this class. This is just for testing purpose. And three files are
>>>quite small: about 8 MB, 5 MB and 3 MB.
>>>
>>>     test.200000.owl: one 'Concept' class and 120000 instances of it
>>>      test.120000.owl: one 'Concept' class and 120000 instances of it
>>>      test.50000.owl: one 'Concept' class and 50000 instances of it
>>>
>>>   On my computer I can load test.50000.owl within 2 minutes,
>>>test.120000 within 5 minutes, but can't load test.200000.owl (I waited
>>>for 30 minutes). My computer has about 1GB memory.
>>>
>>>   If your computer is more powerful than mine, you can use the little
>>>test.cc to generate an ontology with more than 200000 instances of a
>>>class, (you only need to pass a 'number'  you want at the command line
>>>when executing 'a.out' ) and then test agin.
>>>
>>>   If I am not doing something wrong, it looks like Protege doen't
>>>allow a class to have a certain number of instances; otherwise, it
>>>would have difficluty in loading the file. This is true to database
>>>option.
>>>
>>> If you received this email, please kindly let me know.
>>>Thank you
>>>
>>>
>>>
>>>
>>>On 9/7/06, *Mail Delivery System* < MAILER-DAEMON at lists.stanford.edu
>>><mailto:MAILER-DAEMON at lists.stanford.edu>> wrote:
>>>
>>>    This is the Postfix program at host mailman.stanford.edu
>>>    <http://mailman.stanford.edu>.
>>>
>>>    I'm sorry to have to inform you that your message could not
>>>    be delivered to one or more recipients. It's attached below.
>>>
>>>    For further assistance, please send mail to <postmaster>
>>>
>>>    If you do so, please include this problem report. You can
>>>    delete your own text from the attached returned message.
>>>
>>>                            The Postfix program
>>>
>>>    <protege-discussion at mailman.stanford.edu
>>>    <mailto:protege-discussion at mailman.stanford.edu>>: mail forwarding
>>>    loop for
>>>         protege-discussion at mailman.stanford.edu
>>>    <mailto:protege-discussion at mailman.stanford.edu>
>>>
>>>
>>>    Final-Recipient: rfc822; protege-discussion at mailman.stanford.edu
>>>    <mailto:protege-discussion at mailman.stanford.edu>
>>>    Action: failed
>>>    Status: 5.0.0
>>>    Diagnostic-Code: X-Postfix; mail forwarding loop for
>>>        protege-discussion at mailman.stanford.edu
>>>    <mailto:protege-discussion at mailman.stanford.edu>
>>>
>>>
>>>
>>>    ---------- Forwarded message ----------
>>>    From: "Congmin min" < marlonmin at gmail.com
>>>    <mailto:marlonmin at gmail.com>>
>>>    To: protege-discussion at lists.Stanford.EDU
>>>    <mailto:protege-discussion at lists.Stanford.EDU>
>>>    Date: Wed, 6 Sep 2006 14:54:26 -0500
>>>    Subject: [protege-discussion] Number of Instances of one class can
>>>    not be large??
>>>    I am doing some experiments and seems to find another BUG. To
>>>    officially
>>>    report that, I want to ask:
>>>
>>>    Is there any restrictions on the number of instances of one class?
>>>      
>>>
>>The
>>    
>>
>>>    problem I have now is that once the number of instances of one
>>>    class exceeds
>>>    a certain number, Protege failed to load the OWL File, although
>>>    the file is
>>>    pretty small, just several MBs. It is not a memory issue.
>>>
>>>    Thanks for any info.
>>>    _______________________________________________
>>>    protege-discussion mailing list
>>>    protege-discussion at lists.stanford.edu
>>>    <mailto:protege-discussion at lists.stanford.edu>
>>>    https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>>>    <https://mailman.stanford.edu/mailman/listinfo/protege-discussion>
>>>
>>>
>>>
>>>      
>>>
>>_______________________________________________
>>protege-discussion mailing list
>>protege-discussion at lists.stanford.edu
>>https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>>
>>    
>>
>_______________________________________________
>protege-discussion mailing list
>protege-discussion at lists.stanford.edu
>https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>_______________________________________________
>protege-discussion mailing list
>protege-discussion at lists.stanford.edu
>https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
>  
>




More information about the protege-discussion mailing list