# did you provide enough heap to the JVM process too? -> -Xmx3G -Xms3G 
# Did you configure batch.properties to reflect your target store file sizes ? 
(# of nodes x 14bytes for nodes and # of rels x 33 bytes for relationships) ?
-> the total amount for all stores should be total heap - 500MB -> otherwise it 
will swap out to disk which you def. don't want
# 4GB might be a little to small in terms of RAM if there is also other stuff 
running
# make sure your disk is fast enough and configured correctly: 
http://structr.org/blog/neo4j-performance-on-ext4

HTH


Am 19.12.2013 um 00:07 schrieb Puneet Arora <[email protected]>:

> Hi All,
> 
> I am using Neo4j (CSV) Batch Importer 
> (https://github.com/jexp/batch-import/tree/20) to create database from CSV 
> files.
> The data contains 5,115,915 nodes and 113,122,279 edges.
> 
> It takes 128s to load all the Nodes, but a lot of time create edges. 
> Approximately, it creates 10,000,000 edges in about 30-40 mins.
> The edges only contain one property.
> 
> Maybe I am doing something wrong as on GITHUB it says "Importing 41246740 
> Relationships took 170 seconds". 
> I desire to have this speed or anyhow near to it is fine.
> 
> I am running the batch-import on Ubuntu 12.04 64bit, Dual core processor 
> (2.0Ghz each) and a RAM of 4GB.
> I have also done configurations in batch.properties file to increase the 
> relationshipstore.db.mapped_memory = 2500m (But still, no increase in 
> performance)
> 
> Kindly, inform if this is the expected behavior OR how can I increase the 
> speed ?
> 
> Thanks
> PA
> 
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Neo4j" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> For more options, visit https://groups.google.com/groups/opt_out.

-- 
You received this message because you are subscribed to the Google Groups 
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to