Quick Question Regarding eXist performance
|
View:
New views
5 Messages
—
Rating Filter:
Alert me
|
 |
Quick Question Regarding eXist performance
by Kris McGlinn-3
::
Rate this Message:
Hi Everyone,
I am just running some performance testing on eXist. I was wondering if there were any metrics for the times it takes to put resources into the data base? At the moment I seem to be writing to the data base (using putResource) in about one hundredth of a second. The machine is an Intel Core 2 Quad CPU @ 2.66 GHz with 4 GBs RAM. I would be happier if I could keep this down much lower, to perhaps a thousandth of a second or less. Is that about the time I should expect, or can I improve the performance? All help much appreciated, Thanks, Kris. -- Kris McGlinn, MSc. Research Assistant, Ph.D Student Knowledge & Data Engineering Group F.35, Computer Science Department, Trinity College Dublin Kris.McGlinn@... ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Exist-open mailing list Exist-open@... https://lists.sourceforge.net/lists/listinfo/exist-open |
|
|
Re: Quick Question Regarding eXist performance
by Adam Retter-3
::
Rate this Message:
There are almost certainly things that you could try to increase this speed...
Have you disabled indexing (including full-text indexes) - see conf.xml? 2009/11/6 Kris McGlinn <Kris.McGlinn@...>: > Hi Everyone, > > I am just running some performance testing on eXist. I was wondering if > there were any metrics for the times it takes to put resources into the > data base? > > At the moment I seem to be writing to the data base (using putResource) > in about one hundredth of a second. The machine is an Intel Core 2 Quad > CPU @ 2.66 GHz with 4 GBs RAM. I would be happier if I could keep this > down much lower, to perhaps a thousandth of a second or less. Is that > about the time I should expect, or can I improve the performance? > > All help much appreciated, > > Thanks, > > Kris. > > -- > Kris McGlinn, MSc. > Research Assistant, Ph.D Student > Knowledge & Data Engineering Group > F.35, Computer Science Department, Trinity College Dublin > Kris.McGlinn@... > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Exist-open mailing list > Exist-open@... > https://lists.sourceforge.net/lists/listinfo/exist-open > -- Adam Retter eXist Developer { United Kingdom } adam@... irc://irc.freenode.net/existdb ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Exist-open mailing list Exist-open@... https://lists.sourceforge.net/lists/listinfo/exist-open |
|
|
Re: Quick Question Regarding eXist performance
by Peter Crowther
::
Rate this Message:
2009/11/6 Kris McGlinn <Kris.McGlinn@...>
At the moment I seem to be writing to the data base (using putResource) I love the way Intel has managed to brand "CPU speed = computer speed". It's just plain wrong for most problems, and this is particularly obvious with databases. For databases, the limiting factor is almost always disk latency and throughput. In particular, consider a modern 7200rpm desktop or cheap server disk (120 revolutions per second, or c. 8ms per revolution). To write data at an arbitrary location on the disk, it needs to seek (typically 4ms or so) and wait for the sector to come under the disk head (typically another 4ms), for about 8ms per random write. Unless you have write combining/reordering turned on, or you have disk write cache, this limits each physical disk to about 120 random writes per second. This may be most of your problem. Tuning disk I/O can be an interesting task, so here are a few starter questions. How fragmented are your eXist data files on the disk? What is your disk subsystem? How fast are the disks - rotational speed and mean seek time? Is disk write caching enabled all the way through the system, and if so are you aware of the risks you're running with transactional data? Are the disks or the drivers doing write reordering, and if so are you aware of the risks you're running with transactional data? Do you have battery-backed write cache so that a power failure with caching and reordering enabled doesn't run the risk of irretrievably corrupting your database (unless you've deliberately bought a server-class system with BBWC, you almost certainly will not have)? Regards, - Peter ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Exist-open mailing list Exist-open@... https://lists.sourceforge.net/lists/listinfo/exist-open |
|
|
Re: Quick Question Regarding eXist performance
by Kris McGlinn-4
::
Rate this Message:
Hi Adam,
Thanks for that. I changed the following: <index> <fulltext attributes="none" default="false"> <exclude path="/auth"/> </fulltext> </index> It doesn't seem to have made a significant impact on the speed, is there anything else you could recommend? Thanks, Kris. Adam Retter wrote: > There are almost certainly things that you could try to increase this speed... > > Have you disabled indexing (including full-text indexes) - see conf.xml? > > > > 2009/11/6 Kris McGlinn <Kris.McGlinn@...>: > >> Hi Everyone, >> >> I am just running some performance testing on eXist. I was wondering if >> there were any metrics for the times it takes to put resources into the >> data base? >> >> At the moment I seem to be writing to the data base (using putResource) >> in about one hundredth of a second. The machine is an Intel Core 2 Quad >> CPU @ 2.66 GHz with 4 GBs RAM. I would be happier if I could keep this >> down much lower, to perhaps a thousandth of a second or less. Is that >> about the time I should expect, or can I improve the performance? >> >> All help much appreciated, >> >> Thanks, >> >> Kris. >> >> -- >> Kris McGlinn, MSc. >> Research Assistant, Ph.D Student >> Knowledge & Data Engineering Group >> F.35, Computer Science Department, Trinity College Dublin >> Kris.McGlinn@... >> >> >> ------------------------------------------------------------------------------ >> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day >> trial. Simplify your report design, integration and deployment - and focus on >> what you do best, core application coding. Discover what's new with >> Crystal Reports now. http://p.sf.net/sfu/bobj-july >> _______________________________________________ >> Exist-open mailing list >> Exist-open@... >> https://lists.sourceforge.net/lists/listinfo/exist-open >> >> > > > > -- Kris McGlinn, MSc. Research Assistant, Ph.D Student Knowledge & Data Engineering Group F.35, Computer Science Department, Trinity College Dublin Kris.McGlinn@... ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Exist-open mailing list Exist-open@... https://lists.sourceforge.net/lists/listinfo/exist-open |
|
|
Re: Quick Question Regarding eXist performance
by hst
::
Rate this Message:
Kris,
if you have lucene fulltext indexing enabled, the 100msec times you mention are basically what we observed in a small experiment, too (see [1]). However, with no fulltext indexes, performance was much better [2]. That might have to do with increased cache and ram assignment or many other factors.
Best, Heiko --
Heiko Stoermer University of Trento, Italy Dept. of Information Science and Engineering (DISI) http://disi.unitn.it/~stoermer OKKAM id: http://www.okkam.org/entity/ok5f23a5ce-a683-4c4d-ae73-b78cdc17aec1 On Mon, Nov 9, 2009 at 10:09 AM, Kris McGlinn <mcglink@...> wrote: Hi Adam, ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Exist-open mailing list Exist-open@... https://lists.sourceforge.net/lists/listinfo/exist-open |
| Free embeddable forum powered by Nabble | Forum Help |
