Bugs item #1982101, was opened at 2008-06-02 15:14
Message generated for change (Comment added) made by vzzzbx
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1982101&group_id=56967
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: PF/runtime
Group: Pathfinder 0.24
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Wouter Alink (vzzzbx)
Assigned to: Nobody/Anonymous (nobody)
Summary: XQ: slowing down
Initial Comment:
MonetDB/XQuery slows down after each query. It seems to be related to the
complexity of the query. If running a simple query (for example "1+1") 10000
times, the first time it executes in .01 second. Then the execution time
gradually increases. The last time it executes in .05 seconds. Only restarting
mserver seems to reduce the execution time again to .01
The following script illustrates the problem:
n=0;
while [ $n -le 10000 ]
do
if [ $(($n % 100)) -eq 0 ]
then
clear
echo "Iteration $n"
time echo '1+1' | mclient -lxq > /dev/null;
fi
echo '1+1' | mclient -lxq > /dev/null;
n=$(($n + 1))
done
Is this expected behavior? (i know that 10000 times executing 1+1 is not really
a good example, but the same behavior is observed with more complex queries)
i'm using: MonetDB-preJune supersourcetarball (28-may-2008), Fedora Core 8,
64bit-oid
----------------------------------------------------------------------
>Comment By: Wouter Alink (vzzzbx)
Date: 2008-06-04 15:30
Message:
Logged In: YES
user_id=621590
Originator: YES
i tried it again with the nightly build, and again, the same behavior as i
reported earlier occured.
the execution time seems to gradually increase over time.
the following steps were taken:
-installed using "monetdb-install.sh --enable-xquery"
-deleted the dbfarm and log folders
-started mserver with: Mserver --dbinit="module(pathfinder);" --set
xquery_backend=algebra
p.s. i perhaps forgot to mention that i was testing the algebra version,
the behavior does not occur with the mps frontend.
----------------------------------------------------------------------
Comment By: Lefteris Sidirourgos (lsidir)
Date: 2008-06-04 10:37
Message:
Logged In: YES
user_id=1856546
Originator: NO
This is an interesting observation, but I really cant think from where to
start looking for the problem. Is it true that this delay can be an OS
accumulate overhead? For the more complex queries, do you see an increase
of only few milliseconds (e.g. .04 milliseconds) or this delay is
increasing with the increase of the complexity of the query?
Thank you,
lefteris
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 17:18
Message:
Logged In: YES
user_id=572415
Originator: NO
At least with the tiny "1+1" query, I cannot reproduce the problem; time
per query stays at .01 sec:
========
$ n=0 ; while [ $n -le 10000 ] ; do if [ $(($n % 1000)) -eq 0 ] ; then
echo "`date` Iteration $n" ; echo '1+1' | time mclient -lx -t >/dev/null ;
fi ; echo '1+1' | mclient -lx -t >/dev/null ; n=$(($n + 1)) ; done
Tue Jun 3 16:49:16 CEST 2008 Iteration 0
0.00user 0.00system 0:00.01elapsed 26%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:49:28 CEST 2008 Iteration 1000
0.00user 0.00system 0:00.01elapsed 27%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:49:39 CEST 2008 Iteration 2000
0.00user 0.00system 0:00.01elapsed 36%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:49:51 CEST 2008 Iteration 3000
0.00user 0.00system 0:00.01elapsed 30%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:50:02 CEST 2008 Iteration 4000
0.00user 0.00system 0:00.01elapsed 20%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:50:14 CEST 2008 Iteration 5000
0.00user 0.00system 0:00.01elapsed 36%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+578minor)pagefaults 0swaps
Tue Jun 3 16:50:25 CEST 2008 Iteration 6000
0.00user 0.00system 0:00.01elapsed 30%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:50:37 CEST 2008 Iteration 7000
0.00user 0.00system 0:00.01elapsed 10%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:50:49 CEST 2008 Iteration 8000
0.00user 0.00system 0:00.01elapsed 20%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
Tue Jun 3 16:51:00 CEST 2008 Iteration 9000
0.00user 0.00system 0:00.01elapsed 20%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+578minor)pagefaults 0swaps
Tue Jun 3 16:51:12 CEST 2008 Iteration 10000
0.00user 0.00system 0:00.01elapsed 30%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+577minor)pagefaults 0swaps
========
The above log actually represents the 3rd or 4th 10k query on on the same
Mserver session, i.e., I actually executed 30k-40k queries without any
visible performance degradation...
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1982101&group_id=56967
-------------------------------------------------------------------------
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
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
