Hi all,
Thanks to everybody that chimed in!
The large amount of memory is required by our 2 year old LingPipeSpotter
implementation, as it loads into memory a prefix tree with all surface
forms we indexed from Wikipedia. Since then, we have implemented 5-6 other
spotters that do not require much memory to run. Therefore there are three
alternatives in Vinoth's case:
1) use a spotter dictionary compatible with his available RAM (we
distribute it in 4 different sizes)
2) use another spotter implementation (e.g. NESpotter,
OpenNLPChunkerSpotter, OpenNLPNGramSpotter, etc.)
3) implement a disk-backed LingPipeSpotter
The context index is not loaded into main memory, although we've had users
requesting that ability. It is in our feature requests to have a
configuration flag for this. So the size of that index should not impact
your RAM needs (but it will change the time performance of the system).
Cheers,
Pablo
On Sun, Jun 24, 2012 at 1:52 PM, Hector Liu <[email protected]> wrote:
> Hi, Vinoth,
>
> I am glad to help here, but you can get more help if you also send the
> question to the mailing list: [email protected].
> More people will receive your question and they can help too.
>
> I haven't encounter such error, it seemed to me that "*
> java.net.BindException:
> No free port within range*" was the problem.
>
> I suggested you check which process is using the 2222 port and kill it
> (only if you know what the process is doing). Or you can change your
> setting in "server.properties" to another address, which is this line:
>
> org.dbpedia.spotlight.web.rest.uri = http://localhost:2222
>
> Change the port number to a free one.
>
> Best,
> Hector
>
>>
>>>>>>>>>>> On Sun, Jun 24, 2012 at 12:43 AM, Dirk Weissenborn <
>>>>>>>>>>> [email protected]> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hey vinoth,
>>>>>>>>>>>>
>>>>>>>>>>>> it seems you dont have the 16GB RAM you need (Invalid maximum
>>>>>>>>>>>> heap size: -Xmx16g, The specified size exceeds the maximum
>>>>>>>>>>>> representable size.). We are currently working on splitting
>>>>>>>>>>>> our data into various splits, but it is not fully implemented yet.
>>>>>>>>>>>> With
>>>>>>>>>>>> this feature it would be possible to load just a split of the
>>>>>>>>>>>> whole data
>>>>>>>>>>>> into memory (less than 16GB). Anyway until then I guess your
>>>>>>>>>>>> memory space
>>>>>>>>>>>> is not big enough.
>>>>>>>>>>>>
>>>>>>>>>>>> Best
>>>>>>>>>>>> Dirk
>>>>>>>>>>>>
>>>>>>>>>>>>
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Dbp-spotlight-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbp-spotlight-users
