[svg-developers] Re: Performance

Hi Friedhelm, Three seconds instead of one seems too much to me. If you would show us the code somebody might be able to help. Regards Domenico --- In svg-developers@yahoogroups.com, "friedhelmeichin" <[EMAIL PROTECTED]> wrote: > > I am trying to build a scada mmi prototype with svg. For t

[svg-developers] Re: Performance

Hi, I have extracted the essential code snippets. In this case the interval is 1 second but 75% CPU usage (P4/1.5GHz), if I add more controls first cpu usage increaes and later the interval grows too. But if I remove the background image, cpu usage is about 5% I am using javaScript only for te

[svg-developers] Re: Performance

Hi Friedhelm, before I take a closer look I just renamed the 15 id="frame" that re- occur in all groups and it seems to work fine with an image in the bg (athlon 2.2 - cpu 7-12%, 3-9% without your svg running). Question: is your bg image 100x100? When using a 100x100 img all is fine under an

[svg-developers] Re: Performance

Hi Friedhelm, I dont't get any performance problems - and you probably would not want my machine I am using for free :) Your modified sample: http://www.zuccaralloo.de/devgroup/samples/performance.svg modifications: image element: I removed fill and stroke-opacity (never have seen fill being

[svg-developers] Re: Performance

Heiko wrote: > I dont't get any performance problems - and you probably would not > want my machine I am using for free :) You certainly don't get performance problems with a muddy duotone 23.3k jpeg. With some other image like in: http://www.dotuscomus.com/svg/test/performance1.svg cpu usage

[svg-developers] Re: Performance

Hi Nico, I had no idea that this was a picture contest ;-) Anyhow, I guess most members of this group would have prefered to download a 2MB shiny picture sample to - well - just to see the same result. All I can observe is that no matter what size or how many tones the background picture has th

[svg-developers] Re: Performance

Heiko, no problem! > Due to my > observations it even seems that a picture closer to the size of the > screen is easier to handle than a picture that is scaled to fit > (but there I might get on the wrong track). I have also noticed that. Somebody must have an explanation. > I had no idea tha

[svg-developers] Re: Performance

Hi Nico, Heiko and all other guys, thank you for your tips and your discussion. This morning a tried to get a first conclusion of our discussion. 1. The svg modifications made by heiko are interesting but they get no relevant change of the performance result. 2. The kind of the picture has a

[svg-developers] Re: Performance tips

No, I'm not using styles at all, knowing, that they are not so fast as direct properties. Some other ideas, if Mozilla specific, the better... M. --- In svg-developers@yahoogroups.com, <[EMAIL PROTECTED]> wrote: > > Are you using styles in your svg? > > for example > >

[svg-developers] Re: Performance tips

No CSS usage in my game. Some other tips please? --- In svg-developers@yahoogroups.com, <[EMAIL PROTECTED]> wrote: > > Are you using styles in your svg? > > for example > >

[svg-developers] Re: Performance tips

Thx for the tip to sound managers => I'll try them, at least. For the game test, sorry, but I'm not sure on which address I could contact you with mail attachments... M. --- In svg-developers@yahoogroups.com, Erik Dahlström <[EMAIL PROTECTED]> wrote: > > On Thu, 21 Sep 2006 11:55:38 +0200, revelo

[svg-developers] Re: Performance tips

I Phi, glad to read information like this, but in case A, I'm not sure if possible for type of operation I need to do and for B case, you advise to set minimum of attributes to each element? I'm not sure if there is a lot of these cases in my code. But for the A case, my situation is: Background

[svg-developers] Re: Performance tips

After implementing all three optimize attributes to root element and removing of the only one one element with opacity (circle-sun) I could still easily during movement reach 100% of CPU load :-( --- In svg-developers@yahoogroups.com, "Phi Tran" <[EMAIL PROTECTED]> wrote: > > More tips. > - Only u

[svg-developers] Re: Performance tips

My many thanks to all of you trying to help me. There is many hints but I'll try to answer them globally at once there: - I tried to avoid clip-path and it is as one game option to disable. It helps a bit, but not much. My clipping path is done using two primitives - circles; - about javascript its

[svg-developers] Re: Performance tips

"revelonshift" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > Hello there. Is there somebody experienced in > high-perrformance-tweaking of SVG implementation in Mozilla? I did > almost everything I could think about for optimization, but still, my > Submarine close-to-finish game

[svg-developers] Re: Performance tips

I also think that getting infoamtion from DOM on the fly is slowing-down operations, so I've already cached all nodes in variables and using only write - setAttributeNS method. But maybe using directly SVG DOM could speed some time, but I'm not sure how to apply changes to trasform matrix, especial

[svg-developers] Re: Performance tips

Thank you. It works,. I tried the same yesterday without success but just today realize to get initial transform object I have to specifify initialization in SVG like this transform="matrix(1 0 0 1 0 0)" (for the fist time). Then it works. Now I replaced the most often setAttribute() calls and see

[svg-developers] Re: Performance tips

OK, thx, you're right. But since getting matrix is done only once at the initiation phase and during gameplay is only updated, I will probably stay at my previous solution... Anyway .. thx for your point... but, during game I'm often affecting fill attribute and sometimes using rgb color schema. Is

[svg-developers] Re: Performance tips

revelonshift wrote: > OK, thx, you're right. But since getting matrix is done only once at > the initiation phase and during gameplay is only updated, I will > probably stay at my previous solution... > Anyway .. thx for your point... but, during game I'm often affecting > fill attribute and someti

[svg-developers] Re: Performance problem

--- In svg-developers@yahoogroups.com, "slavaraitsin" <[EMAIL PROTECTED]> wrote: > > Hi, > I'm working on a rather big project using SVG as rendering engine. > We are facing mayor performance problems when using removeChild java > script method to remove or replace a subtree of the SVG DOM. > D

[svg-developers] Re: Performance problem

> Try using the internal the ASV javascript engine as it is based on mozilla's > js...It will be interesting to hear if there is any improvement. > > http://wiki.svg.org/index.php/AsvJavaScriptEngine I did it testing on IE ASV engine works (removeChild) even sligtly slower than internal IE ja

Re: [svg-developers] Re: Performance

As a wild guess, set the pointer-events property on the image to fill or all. If it is being treated as a rectangle then the implementation can take some shortcuts (not have to worry about transparency). This will save a couple of steps in the re-render. But this is just a guess. All the best,

[svg-developers] Re: Performance question

Hi Julie, can you post an example to see how you structured the file? Also which SVG viewer did you try? Also on which OS did you try, with how much memory and CPU power available? But it might well be that 36000 circles are too much for the current SVG viewers. The currently existing viewers (A

[svg-developers] Re: Performance question

"jgfa92004" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > I have a svg file containing 36000 circles That's a lot of circles! on a 1024x768 they'd all have to have a radius of 2 pixels so you could see them all, at 2px radius, there's no way you could see they are circles. Do

[svg-developers] Re: Performance question

Hi Julie, I still don't understand if you generate the circles in the client, using Javascript or on the server. In the first case, that might be one of the reasons why it is very slow. If you generate the circles on the server, than you can as well aggregate the circles. I had a similar pro

[svg-developers] Re: Performance question

have you tried to load the file in Batik and see what part takes the longest? In Batik you get messages in the status line, telling when the file was loaded, when the scripts are parsed and when rendering starts. Is it the rendering that takes that long or the parsing of the file? If you use

Re: [svg-developers] Re: Performance tips

revelonshift a écrit : > No CSS usage in my game. Some other tips please? avoid transparency like the plague. ___ Découvrez un nouveau moyen de poser toutes vos questions quelque soit le s

Re: [svg-developers] Re: Performance tips

In my experience I use this scheme and it helps. I use multiple layers specially those that serve as the background. Then I only "redraw" the layer that need to change. In this case I do not force the player to scan/redraw the whole SVG. Note: Every time updating the player does not know which and

Re: [svg-developers] Re: Performance tips

More tips. - Only use Semi-transparent or Semi-opacity when it is in real need. A pixel is processed multiple times faster if it not Semi-opacity.(this is one of the big time consuming work horse). - One more tip that you should have in mind in creating the composition of your SVG file/layer. If yo

Re: [svg-developers] Re: Performance tips

One more tip: Void to use clip-path at all if you can. This is one of the worst scenarios a player should face. It has to generate a path, save it somewhere a background, Then compare every of the backgound/foreground pixel by pixel to see if it is in/out the clip then to decide if it need to pain

Re: [svg-developers] Re: Performance tips

Revelon, Optimisation is often browser-specific, beyond the tried and true methods such as the good ones which Phi Tran offered. Have you considered looking at whether it is your svg code or your scripting which is eating CPU cycles? Complex javascript is resource consuming if there are no br

Re: [svg-developers] Re: Performance tips

As system developer and 100% CPU. Can You as application developer do anything about it? the answer is: Both Yes and No. NO- 100% CPU only means that the player/or system code go into a loop and not yielding CPU to other thread. In this case you can not do anything. Unless the player allowing yo

Re: [svg-developers] Re: Performance tips

On 9/21/06 7:29 AM, revelonshift wrote: > No, I'm not using styles at all, knowing, that they are not so fast as > direct properties. Some other ideas, if Mozilla specific, the better... There will be no performance difference between style and property attributes in Mozilla SVG, as they are both

Re: [svg-developers] Re: Performance tips

On 9/21/06 10:18 AM, T Rowley wrote: > Major items to keep an eye out for performance in Firefox 1.5/2.0 SVG > are group opacity ("fill-opacity" is fine), clipPaths, and complex paths > (the current cairo tesselator has poor scalability). All of these > should improve in Firefox 3.0. A little

Re: [svg-developers] Re: Performance tips

Lately I am very careful in what I will say specially after ADOBE's announcement. For not to be misunderstood as a "cheap-shot" to promote my SVG/Zip . . . . Let me make my point on this thread. I only trying to help to save some CPU instructions here and there; Base on my background and study of

Re: [svg-developers] Re: Performance tips

On 9/21/06 10:42 AM, Phi Tran wrote: > Lately I am very careful in what I will say specially after ADOBE's > announcement. For not to be misunderstood as a "cheap-shot" to promote my > SVG/Zip . . . . > > Let me make my point on this thread. I only trying to help to save some CPU > instructions he

Re: [svg-developers] Re: Performance tips

On 9/21/06, Jim Ley <[EMAIL PROTECTED]> wrote: > > > "revelonshift" <[EMAIL PROTECTED] > wrote > in message > news:[EMAIL PROTECTED] . > > Hello there. Is there somebody experienced in > > high-perrformance-tweaking of SVG implementation in Mozilla? I did > > almost everything I could think about f

Re: [svg-developers] Re: Performance tips

On 9/21/06 11:27 AM, Jim Ley wrote: > The biggest Mozilla speed improvement in scripting will be to not use > getElementById() but to cache every single reference in a seperate variable > of your own. currently mozilla re-looks up every node in the DOM for XML > documents, and it's very, very s

Re: [svg-developers] Re: Performance tips

Ouch. Good to know. Thanks, Jim. On Thursday 21 September 2006 18:27, Jim Ley wrote: > The biggest Mozilla speed improvement in scripting will be to not use > getElementById() but to cache every single reference in a seperate variable > of your own.  currently mozilla re-looks up every node in th

Re: [svg-developers] Re: Performance tips

On 9/21/06, Phi Tran <[EMAIL PROTECTED]> wrote: > > > > On 9/21/06, Jim Ley <[EMAIL PROTECTED]> wrote: > > > > > > "revelonshift" <[EMAIL PROTECTED] > wrote > > in message > > news:[EMAIL PROTECTED] . > > > Hello there. Is there somebody experienced in > > > high-perrformance-tweaking of SVG implem

Re: [svg-developers] Re: Performance tips

On Thu, 21 Sep 2006 19:28:16 +0200, revelonshift <[EMAIL PROTECTED]> wrote: > I also think that getting infoamtion from DOM on the fly is > slowing-down operations, so I've already cached all nodes in variables > and using only write - setAttributeNS method. > But maybe using directly SVG DOM co

Re: [svg-developers] Re: Performance tips

Den 2006-09-22 11:24:53 skrev revelonshift <[EMAIL PROTECTED]>: > I tried the same yesterday without success but > just today realize to get initial transform object I have to specifify > initialization in SVG like this transform="matrix(1 0 0 1 0 0)" (for > the fist time). You could use the DOM

Re: [svg-developers] Re: Performance tips

On Fri, 22 Sep 2006 12:06:20 +0200, revelonshift <[EMAIL PROTECTED]> wrote: > OK, thx, you're right. But since getting matrix is done only once at > the initiation phase and during gameplay is only updated, I will > probably stay at my previous solution... > Anyway .. thx for your point... but,

Re: [svg-developers] Re: Performance problem

Hi, Can you let me know how it is done on Mozilla? Thanks ggb2g <[EMAIL PROTECTED]> wrote: --- In svg-developers@yahoogroups.com, "slavaraitsin" <[EMAIL PROTECTED]> wrote: > > Hi, > I'm working on a rather big project using SVG as rendering engine. > We are facing mayor performance problems

RE: [svg-developers] Re: Performance problem

>>I am planning to test in Mozilla SVG enable version to see if I have the same issue when not using a plugin.<< Try using the internal the ASV javascript engine as it is based on mozilla's js...It will be interesting to hear if there is any improvement. http://wiki.svg.org/index.php/AsvJavaScri

Re: [svg-developers] Re: Performance problem

I'm curious if you can just avoid calling removeChild. Here are some alternatives: - set the display attribute to "none" this will make it not visible - call appendChild on another element this will remove from current location and put in the new location. You could make a document fragment, cal

RE: [svg-developers] Re: Performance question

Thanks for your quick answer. In my svg file, I have first a javascript section that contains an array that have the same nb of lines (36000) Then here is a part of my svg file section where circles are drawn : ... I use the Adobe SVG viewer on IE6, on Windows XP, 2.6 GHz

Re: [svg-developers] Re: Performance question

hi julie, there is no real workaround. ("trop c'est trop.") some possibilities although: - switch of anti-aliasing - try to render rectangles rather than circles. if they are that small that won't make any difference. i did some tests a year ago with randomly generated basic shapes. in order

Re: [svg-developers] Re: Performance question

What do you mean by "switch of anti-aliasing" ? Concerning your suggestion of create rectangles instead of circles, why not... I'll test it. I don't understand neither your last suggestion. What I do is to generate the file using Batik, so each circle is written inside a loop (creating the circle

Re: [svg-developers] Re: Performance question

hi julie, >What do you mean by "switch of anti-aliasing" ? > > check out ASV supports that. >Concerning your suggestion of create rectangles instead of circles, why not... >I'll test it. >I don't understand neither your last

RE: [svg-developers] Re: Performance question

I create the circles on the server, so when the svg is displayed, the circles are already written in the file. But what you say is important because I was wondering if it could be faster if I'd create the circles by using a javascript function when loading the document... So the response seems t

[svg-developers] Re: Performance problem while setAttribute("points") on polyline

You're right, the asv rendering engine is dead slow. Try to lower demands on viewer: -set xxx-rendering="optimizeSpeed", at least shape- rendering="crispEdges" (probably you'll know already;-) -asv antialiasing is very expensive, set defaultAntialias="false", or use .setDefaultAntialias

[svg-developers] Re: Performance problem while setAttribute("points") on polyline

: +92 42 5427001 (3 lines) > Fax: +92 42 5411724 > Web: www.intechww.com > > > > "welkerpaul" <[EMAIL PROTECTED]> > 02/02/2005 06:52 PM > Please respond to > svg-developers@yahoogroups.com > > > To > svg-developers@yahoogroups.c

[svg-developers] Re: Performance differences : IE/ASV3 and Op9 vs FF2.0

--- In svg-developers@yahoogroups.com, "simonshutter" <[EMAIL PROTECTED]> wrote: > > I have to change the attributes of a large number of elements. I use > the code below: > > var i=0; > var elm; > var r=document.getElementById('test'); > var colR=r.getElementsByTagNameNS ('http://www.w3.org/2

[svg-developers] Re: Performance differences : IE/ASV3 and Op9 vs FF2.0

> The loop is correct. FF should perform 2 to 10 times slower, between > 30 and 150 ms, nothing like 115682 msec. The problem is elsewhere. > > Domenico I think I've identified the cause of the performance difference and it is the "use" element: If you load the example below you can see th

[svg-developers] Re: Performance differences : IE/ASV3 and Op9 vs FF2.0

--- In svg-developers@yahoogroups.com, T Rowley <[EMAIL PROTECTED]> wrote: > > That will have poor performance in Firefox due to they way we currently > implement , which is to do an anonymous clone of the indicated > subtree at the point of the . Any modifications to the original > content

Re: [svg-developers] Re: Performance problem while setAttribute("points") on polyline

Cell: 03334241582 Voice: +92 42 5427001 (3 lines) Fax: +92 42 5411724 Web: www.intechww.com "welkerpaul" <[EMAIL PROTECTED]> 02/02/2005 06:52 PM Please respond to svg-developers@yahoogroups.com To svg-developers@yahoogroups.com cc Subject [svg-developers] Re: Perf

Re: [svg-developers] Re: Performance differences : IE/ASV3 and Op9 vs FF2.0

On 3/27/07 2:49 PM, simonshutter wrote: >> The loop is correct. FF should perform 2 to 10 times slower, > between >> 30 and 150 ms, nothing like 115682 msec. The problem is elsewhere. > > I think I've identified the cause of the performance difference and > it is the "use" element: > > > > I