Hi Alexander,

I presume 50kk in 50 million ?

You indicated having allocate all you memory for use by the Virtuoso Server 
whereas it is recommended about 60% of memory is allocated as detailed at:

        
http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtRDFPerformanceTuning

Also, are you running the Virtuoso RDF Bulk loader scripts detailed at:

        
http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtBulkRDFLoaderScript

as these are what we used for loading the billion plus datasets loading into 
some of the endpoints we use. These script auto set log_enable mode prior to 
upload. Note you can also run multiple rdf_runloader() scripts to perform 
parallel loading of the datasets for better performance. I’ve had upwards for  
300 million triple uploads per hour on an 8 core machine with 72GB of memory 
running 8 instances of the rdf_runloader() scripts in parallel ( one for each 
core).

Best Regards
Hugh Williams
Professional Services
OpenLink Software
Web: http://www.openlinksw.com
Support: http://support.openlinksw.com
Forums: http://boards.openlinksw.com/support
Twitter: http://twitter.com/OpenLink

On 26 Jun 2011, at 17:00, Alexander Sidorov wrote:

> Hello!
> 
> What does influence on the bulk dataset uploading performance? I need to 
> upload about 50kk triples (quads) to the server with outdated hardware and 
> 1gb of memory. I have set MaxCheckpointRemap, NumberOfBuffers and 
> MaxDirtyBuffers to take all available memory but uploading goes very slow, 
> about 100k triples per hour (I am not even sure whether memory size plays 
> significant role in dataset uploading process). Are there any other points to 
> improve uploading performance? Does log_enable play role here (like in graph 
> deletion)?
> 
> Regards,
> Alexander
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security 
> threats, fraudulent activity, and more. Splunk takes this data and makes 
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2d-c2_______________________________________________
> Virtuoso-users mailing list
> Virtuoso-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/virtuoso-users


Reply via email to