Re: RDF DB Shootout: Preliminary Results

From: Ian Wilson <Ian.Wilson_at_uchsc.edu>
Date: Fri, 13 Jan 2006 09:54:52 -0700

Hi,

I'm interested in the memory based benchmarks. As mentioned
previously on the list, I'd be happy to help with testing the
memory stores in a larger memory space -IBM, 64 GB, 8
processor SMP. I have to schedule time on this machine, but it
is otherwise available. I'm using nmon and rrdtool to generate
statistics - but will take a look at the new test harness
code, and see what I come up with.

I'm primarily working with the Uniprot data set (extracts of
this data set, actually, because I've been unable to load the
full data set). Any reason why this would not be a good 'real
world' data set to include in your tests? It is large, but can
easily be split into smaller chunks.

On a related topic, is anyone aware of efforts to manage a
triple store across a cluster - e.g. using a column store
approach. I noticed the recent c-store project at csail
released code, but have not had a chance to dig into it yet.
Google's BigTable project takes a similar approach.

Ian







Arjohn Kampman said the following on 1/13/2006 2:44 AM:
> Vineet Sinha wrote:
>> I will run the tests again in about 2-3 weeks and post any updates.
>> Also, just to make more apples-to-apples comparison, I will also
>> adding both Jena-MySQL and Sesame-MySQL to the results.
>>
>> If there are any other suggestions/contributions please let me know in
>> the next week or so, and I will be able to push things out in one go.
>
> Hi Vineet,
>
> Considering that you are using "only" 1 million statements in your
> benchmark (right?), you could consider to include Sesame-memory in the
> results too. This should fit easily in the 1GB RAM of your machine and
> it will make the Sesame-part of the picture complete.
>
> --
> Arjohn
>
Received on Fri Jan 13 2006 - 16:54:56 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:18 EDT