Hugh,
it is 64bits:
virtuoso-t: ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for
GNU/Linux 2.4.0, dynamically linked (uses shared libs), stripped
so, we will wait until 5.0.7 is out there...
Erick
Hugh Williams wrote:
On 19 Jun 2008, at 16:00, Erick Antezana wrote:
Hi Hugh,
please see below:
Hugh Williams wrote:
Hi Erick,
On 17 Jun 2008, at 09:11, Erick Antezana wrote:
Hi Hugh,
There was only one defined index on that table (RDF_QUAD_OGPS). I have
added the following:
create index RDF_QUAD_GSPO on DB.DBA.RDF_QUAD (G,S,P,O);
create bitmap index RDF_QUAD_POGS on DB.DBA.RDF_QUAD (P,O,G,S);
create bitmap index RDF_QUAD_PSOG on DB.DBA.RDF_QUAD (P,S,O,G);
create index RDF_QUAD_PSGO on DB.DBA.RDF_QUAD (P,S,G,O);
and the queries are still taking a lot of time although there seems to
be faster in some cases.
[Hugh] OK, so some improvement but still performance issues. Looking
through previous emails I note your are running on Linux but is in
unclear as to whether you are on 32 or 64 bit Linux, thus can you
please confirm. Note that running the file command on your virtuoso-t
binary should confirm whether it is 32 or 64 bit. The reason I ask is
that 32bit Linux would have a limit on of about 4GB on addressable
memory, so if you are using a 32bit binary you would be limited by
this, whereas if you are running a 64bit binary the addressable memory
is of the order of 16 ExtraBytes or whatever limit the OS/hardware has
imposed on the system. So, if the 16GB or intended 32GB of memory on
your machine is to be taken advantage of you need to be running a
64bit binary, thus we need to confirm if this is the case. Assuming
you compiled the binary yourself we would also be interested in the
compiler flags used for your build to see if these are optimal.
it is a 64 bit Linux. And it was compiled with no special flags, just
the typical:
configure
make
make check
make install
would you recommend to compile it like this:
CFLAGS="-O2 -m64"
export CFLAGS
./configure ...
?
[Hugh] Those look reasonable, but I would still like to confirm that
the resultant binary built is actually 64bit, thus can you run the
following from your bin directory of the VOS installation and provide
the output:
file virtuoso-t
we are running version 5.0.3. It seems that on the 5.0.6 there was some
optimization work on the RDF side right? Would you also recommend to go
for 5.0.6? For when is the next release foreseen?
[Hugh] OK, I thought you had already updated to 5.0.6 as their have
been RDF optimizations in this release. A new VOS 5.0.7 release was/is
due out this week and will most probably be available either
tomorrow(Friday) or Monday now I would think. So it would be best to
hold fire for this in my opinion.
Regards
Hugh
Erick
In http://docs.openlinksw.com/virtuoso/rdfperformancetuning.html
it is
recommended to have such similar indexes. I was not able to create the
ones with partitions since it seems that the partition option only
works
when one has a cluster? "...If clustering is not enabled, partitioning
can still be declared but it will have no effect..." (c.f.
http://docs.openlinksw.com/virtuoso/clusteroperationpart.html). Is
virtuoso V5 capable of dealing with that? If so, how should I do it?
I tried to follow the commands about the DBpedia benchmark
(http://docs.openlinksw.com/virtuoso/rdfperformancetuning.html)
however,
after creating table r2 ( create table r2 (g iri_id, s iri_id, p
iri_id, o any, primary key (s, p, o, g)) ) then I tried to alter its
index (alter index R2 on R2 partition (s int (0hexffff00));) I got:
SQL> create table r2 (g iri_id, s iri_id, p iri_id, o any, primary key
(s, p, o, g))
Type the rest of statement, end with a semicolon (;)> ;
Done. -- 122 msec.
SQL> alter index R2 on R2 partition (s int (0hexffff00));
*** Error 37000: [Virtuoso Driver][Virtuoso Server]SQ074: Line 1:
syntax
error at 'index' before 'R2'
at line 3 of Top-Level:
alter index R2 on R2 partition (s int (0hexffff00))
SQL>
[Hugh] This commands are indeed for the Virtuoso Clustered Server
version which is not currently available and thus do not work with the
current version, and thus have been removed from the online
documentation.
Regards
Hugh
Regards,
Erick
--==================================================================
Erick Antezana http://www.cellcycleontology.org
PhD student
Tel:+32 (0)9 331 38 24 fax:+32 (0)9 3313809
VIB Department of Plant Systems Biology, Ghent University
Technologiepark 927, 9052 Gent, BELGIUM
er...@psb.ugent.be http://www.psb.ugent.be/~erant
==================================================================
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Virtuoso-users mailing list
Virtuoso-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/virtuoso-users
--
==================================================================
Erick Antezana http://www.cellcycleontology.org
PhD student
Tel:+32 (0)9 331 38 24 fax:+32 (0)9 3313809
VIB Department of Plant Systems Biology, Ghent University
Technologiepark 927, 9052 Gent, BELGIUM
er...@psb.ugent.be http://www.psb.ugent.be/~erant
==================================================================
