W3C home > Mailing lists > Public > semantic-web@w3.org > June 2013

RE: BUBM: explore and update

From: Peter Boncz <P.Boncz@cwi.nl>
Date: Thu, 13 Jun 2013 22:44:19 +0200 (CEST)
To: "'Jeremy J Carroll'" <jjc@syapse.com>, <semantic-web@w3.org>
Message-ID: <003001ce6876$be302920$3a907b60$@nl>
Dear Jeremy,

This is a very good question, without - regrettably - a good answer.

Our benchmarking exercise initially focused in getting the Business
Intelligence use case tested. When we started, this was a challenges as
this had not been achieved yet. For completeness, we ran also the explore
workload, but forgot about explore+update.

We plan to rerun the benchmark this fall, we will actually try to scale up
to 500 billion triples. At that moment, we will also run the
explore+update.

Thanks,

Peter Boncz



-----Original Message-----
From: Jeremy J Carroll [mailto:jjc@syapse.com] 
Sent: Thursday, June 13, 2013 7:12 PM
To: semantic-web@w3.org
Subject: BUBM: explore and update


Hi

I have been looking at BUBM as a guide to a triple store decision, and I
have been concentrating on the recent V7 version

http://wifo5-03.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/res
ults/V7

I note that the older (out of date) V6, has a section that is missing from
V7

http://wifo5-03.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/res
ults/V6/#resultsExploreAndUpdate

which seems to better reflect our intended usage pattern

Anyone have any insight into: why this mix was not run in V7 and/or how
the triple stores fare with a mix between update and queries

Jeremy


Principal Architect
Syapse, Inc.





-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3345 / Virus Database: 3199/6404 - Release Date: 06/12/13
Received on Friday, 14 June 2013 14:06:51 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:45:33 UTC