Amiga.org

Amiga computer related discussion => Amiga Software Issues and Discussion => Topic started by: 0amigan0 on December 19, 2009, 09:37:15 AM

Title: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 19, 2009, 09:37:15 AM
If I don't go wrong, you have the sources of OWB (68k) made by joerg strohmayer.
You keep saying it is un-optimized, that it doesn't have threads, etc...
Well, improve it, give it threads; after you have "optimized it", someone else might give it a full "native" MUI interface.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 19, 2009, 12:10:33 PM
Quote from: 0amigan0;534445
If I don't go wrong, you have the sources of OWB (68k) made by joerg strohmayer.
You keep saying it is un-optimized, that it doesn't have threads, etc...
Well, improve it, give it threads; after you have "optimized it", someone else might give it a full "native" MUI interface.


The source is outdate and OWB core is change much since some years.If you wish i can send you the source(write PN) or ask Jörg.but you can nothing do with it.its way toooo slow and too old.the OWB team have also speedup it

on AROS can see, OWB is slow, i dont believe it is faster on MOS or OS4, because Cairo cost too more performance.,OWB have no diskcache.On netsurf they are working on diskcache.

Its always good to wait instead spend much work, maybe the red versus blue war end in 1-2 years and working together is better possible of the few existing amiga devs

I need not sell a OS more by have a good browser, so i need not spend much time, i can wait some years so things go faster and browers get better portable and chrome enhance.

Sooner or later i buy a 4 core I5 or I7 then compiling get lots faster, then maybe at least the compile time of this slow C++ monsters is acceptable and i have fun to do it
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 19, 2009, 01:02:03 PM
Quote from: bernd_afa;534456
....the OWB team have also speedup it....


Can't u take THAT source then, and make a new port for 68k ?

Quote


I need not sell a OS more by have a good browser, so i need not spend much time, i can wait some years so things go faster and browers get better portable and chrome enhance.

Sooner or later i buy a 4 core I5 or I7 then compiling get lots faster, then maybe at least the compile time of this slow C++ monsters is acceptable and i have fun to do it


Maybe someone else can do it, Artur perhaps ??
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 19, 2009, 01:29:34 PM
Quote from: bernd_afa;534456
The source is outdate and OWB core is change much since some years.If you wish i can send you the source(write PN) or ask Jörg.but you can nothing do with it.its way toooo slow and too old.the OWB team have also speedup it

on AROS can see, OWB is slow, i dont believe it is faster on MOS or OS4, because Cairo cost too more performance.,OWB have no diskcache.On netsurf they are working on diskcache.

Its always good to wait instead spend much work, maybe the red versus blue war end in 1-2 years and working together is better possible of the few existing amiga devs

I need not sell a OS more by have a good browser, so i need not spend much time, i can wait some years so things go faster and browers get better portable and chrome enhance.

Sooner or later i buy a 4 core I5 or I7 then compiling get lots faster, then maybe at least the compile time of this slow C++ monsters is acceptable and i have fun to do it

Sorry, but this is mostly wrong. Sure WebKit isn't as fast as ibrowse (what a surprise), but it's still *much* faster than NetSurf on the same machine, actually.
And my OWB port (and also OS4 one) is also *much* faster to scroll on my Peg2 than on a Xeon 2.5GHz machine under linux with the plain OWB SDL version. This isn't surprising, considering the SDL implementation doesn't implement a "smart" scroll method at all. As the AROS version doesn't implement that scroll method either and that blitting isn't exactly very fast on AROS, it's normal you found it slow, but it can be much faster.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 19, 2009, 02:40:20 PM
Quote from: Fab;534463
Sorry, but this is mostly wrong. Sure WebKit isn't as fast as ibrowse (what a surprise), but it's still *much* faster than NetSurf on the same machine, actually.
And my OWB port (and also OS4 one) is also *much* faster to scroll on my Peg2 than on a Xeon 2.5GHz machine under linux with the plain OWB SDL version. This isn't surprising, considering the SDL implementation doesn't implement a "smart" scroll method at all. As the AROS version doesn't implement that scroll method either and that blitting isn't exactly very fast on AROS, it's normal you found it slow, but it can be much faster.

The only bottleneck would then be the high compilation time.

@Fab:

How long would it take to cross-compile your version on a *very* powerful Windows machine ?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 19, 2009, 03:09:36 PM
Quote from: 0amigan0;534469
The only bottleneck would then be the high compilation time.

@Fab:

How long would it take to cross-compile your version on a *very* powerful Windows machine ?


On that Xeon PC/Linux I use for crosscompilation, it needs about 30 minutes to compile OWB from scratch (a bit more when SVG is enabled). On my Pegasos2, I tried it once, and it needed 6 hours (but gcc I/O is better handled on Linux, anyway, so it's not surprising to see such a difference). :)

Anyway, on a very recent machine, it would take 15-20 minutes, i guess. I can't comment about Windows, though.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: alexh on December 19, 2009, 07:32:18 PM
If you want to access the web from your Amiga and you have another PC on the network (Mac, Linux, Windows etc.) there are tools that will let a networked Amiga open a web-browser on those machines and display it on the Amiga screen.

With the Amiga hard drives mounted via SAMBA on the host machine (so you can save directly to the Amiga) it looks and feels almost like a native browser only much faster and upto date.

Just google for Amiga & VNC or Remote Desktop.

I have a HP NAS drive (EX470) running Windows Server 2003 which is always on and I click on an icon on the Amiga and a remote FireFox 3.5.6 window opens. Because it is remote everything just works like it would on a PC, Flash, Shockwave, Java etc.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 19, 2009, 07:38:01 PM
Quote from: alexh;534490
If you want to access the web from your Amiga and you have another PC on the network (Mac, Linux, Windows etc.) there are tools that will let a networked Amiga open a web-browser on those machines and display it on the Amiga screen.

With the Amiga hard drives mounted via SAMBA on the host machine (so you can save directly to the Amiga) it looks and feels almost like a native browser only much faster and upto date.

Just google for Amiga & VNC or Remote Desktop.

I have a HP NAS drive (EX470) running Windows Server 2003 which is always on and I click on an icon on the Amiga and a remote FireFox 3.5.6 window opens. Because it is remote everything just works like it would on a PC, Flash, Shockwave, Java etc.

That's too slow for my taste.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 20, 2009, 10:23:30 AM
Quote from: Fab;534463
Sorry, but this is mostly wrong. Sure WebKit isn't as fast as ibrowse (what a surprise), but it's still *much* faster than NetSurf on the same machine


If OWB is faster, then wy netsurf is a google summer of Code projecrt since 2 years and there is no OWB Version in a Linux distri ?.the big disadvantage of netsurf is currently missung java script but its in Linux Distri.

netsurf is in Linux distri.Wy not OWB ?

http://www.drobe.co.uk/article.php?id=2041

maybe the time of render of a complete page is slower or same as OWB, but thats not the important.The important is to get the page as fast as possible vissible and can scroll and click on links.

netsurf have some settings for number of fetchers and the time it refresh.Here users can tweak in config file and get better results on

>On that Xeon PC/Linux I use for crosscompilation, it needs about 30 minutes to compile >OWB from scratch

That sound good.You have 2 core and how many GHZ your system have ?

then i can assume that a Core I7 with 2,66 GHZ can maybe compile it in 12 minutes ?

I have too some time ago compile OWB on my AMD64 3000+
I have compile it 70% sucessfull, but most demotivation was that every time when i type make, the build system check for about over 1 minute to find out on what files it should continue to compile and then it give me a compile error when something is wrong.my ram was enough.

and for that i have not the patience.In this time a netsurf rebuild work and when i change some files in netsurf, a new version can get in 5 sec and a compile error in this file can see in 2 sec.
maybe that slowdown is cygwin relatet ?
when you change a file in OWB on your Linux System, how many seconds it need to detect what file need compile.

>And my OWB port (and also OS4 one) is also *much* faster to scroll on my Peg2 than on >a Xeon 2.5GHz machine under linux with the plain OWB SDL version.

Yes that i believe, but the Reason is, because SDL not support good OWB.OWB use SW_SURFACE and so scroll is done in CPU.

netsurf too use in old versions cpu for scroll, but newest source Version use now SDL to scroll and on HW surface it use the blitter.and as can see in netsurf thread here, its lots faster now when the Pixel format fit ok.

here is the source can easy see  in the diff output

http://source.netsurf-browser.org/trunk/libnsfb/src/frontend_sdl.c?r1=9719&r2=9720
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 20, 2009, 10:49:24 AM
Quote from: bernd_afa;534553
If OWB is faster, then wy netsurf is a google summer of Code projecrt since 2 years and there is no OWB Version in a Linux distri ?.the big disadvantage of netsurf is currently missung java script but its in Linux Distri.

netsurf is in Linux distri.Wy not OWB ?

Why would it be any indication of anything whether something is in specific linux distribution or not?

That's just plain silly.

Quote

maybe the time of render of a complete page is slower or same as OWB, but thats not the important.The important is to get the page as fast as possible vissible and can scroll and click on links.

That's your opinion. I don't agree. Personally I like to have working javascript. I like to have snappy scrolling, too.

Quote
netsurf have some settings for number of fetchers and the time it refresh.Here users can tweak in config file and get better results on

Such things should be automatic. The end user should not be forced to bother with such things.

Quote
I have too some time ago compile OWB on my AMD64 3000+
I have compile it 70% sucessfull, but most demotivation was that every time when i type make, the build system check for about over 1 minute to find out on what files it should continue to compile and then it give me a compile error when something is wrong.my ram was enough.

and for that i have not the patience.

Seriously? You really need to find something else to do than program then.

Quote
In this time a netsurf rebuild work and when i change some files in netsurf, a new version can get in 5 sec and a compile error in this file can see in 2 sec.
maybe that slowdown is cygwin relatet ?
when you change a file in OWB on your Linux System, how many seconds it need to detect what file need compile.

How on earth would this matter at all? The build is automatic anyway, why would you need to know which files need to be compiled? I just don't get what your problem here is.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 20, 2009, 11:01:25 AM
@bernd & Fab (or others):

A proposal:

Bernd, you can start cross-compiling all the libraries required by OWB; it's certainly less work for you, then Fab or another guy (Artur? ), can cross-compile Owb.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 20, 2009, 11:21:44 AM
Quote from: Piru;534554
Why would it be any indication of anything whether something is in specific linux distribution or not?



thats show its intresting for users to use.The fact that OWB is not available in enhanced way for other than MOS OS4 /AROS and is also not listet on wikipedia, do not motivate me for OWB.The only running OWB i can see is on AROS.and here it is 2-3* slower to show the page as soon as possible.

there are lots of browser out you can see too, maybe there are some better browsers here,and thats the reason wy there is no OWB Version in distri.

http://en.wikipedia.org/wiki/Web_browser

I dont want spend much work for a better than nothing solution, i want the best solution, so that when i surf on winuae its same fast as on firefox native.and that is possible with netsurf.and when they add the filecache, then it go faster.

sure java script is miss, and OWB is for MOS OS4 AROS users great, because better wait 5 sec instead of cant show page.but when i see that the page is show on 2 sec(the page layout is done and can scroll) on my windows box with all browsers, i think its not usefull to spend lots work in OWB before the speed get same as other browsers.

>That's your opinion. I don't agree. Personally I like to have working javascript.

I like too have Java script and problem of later show on OWB have nothing to do with Java script.I deactivate it for test in AROS OWB, but get not faster.

>Such things should be automatic. The end user should not be forced to bother with such >things.

but i dont see if that is on OWB and other browsers do something automatic.current browsers are design that they run ok on a System with at least 600 MHZ.and do a rerender every 0.5 sec

When a system is slower then the rerender should be not so often, to slow not too much down.

often rerender increase the full Page load time.
and here its clear wy OWB can show a page faster complete as other browsers.because OWB show the page the first time later as other browsers.

but its always more usefull to show a page as soon as the full text is load and all layout data is here.

then a user can begin read the page, scroll, or click on links.
so in praxis the browser is faster to use, even if the full page load(upto the last promotion banner is load)take longer.

I like want a browse rthat show a page first after 3 sec and if he load the page in 12 sec full i think its lots better as a browser that show the page first after 6 sec and show full in 9 sec.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 20, 2009, 11:45:17 AM
>Bernd, you can start cross-compiling all the libraries required by OWB; it's certainly less >work for you, then Fab or another guy (Artur? ), can cross-compile Owb.

If somebody want port it, i help.
but i do no work, when nobody want do something.he can also try himself to compile Cairo.
 
because software grow and when its really need and begin then too port, you get a more upto date version.

and BTW, somebody who is intresting on Port to 68k,  have fab and me ask about OWB, he install now amidevcpp and want it do in cygwin.i send him my full includes compilers with all libs i have.

and here are most libs in OWB need.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 20, 2009, 12:27:12 PM
Quote from: bernd_afa;534557
thats show its intresting for users to use.
Huh? So you seriously think just because there's software X in linux distros it somehow measures popularity of that said software?

Quote
The fact that OWB is not available in enhanced way for other than MOS OS4 /AROS and is also not listet on wikipedia, do not motivate me for OWB.
But you do know better than that. Instead you just play stupid.

Quote
The only running OWB i can see is on AROS.and here it is 2-3* slower to show the page as soon as possible.
OWB is faster than Netsurf here. The reasons for AROS OWB slowness were explained already.

Quote
I dont want spend much work for a better than nothing solution, i want the best solution, so that when i surf on winuae its same fast as on firefox native.and that is possible with netsurf.and when they add the filecache, then it go faster.
Best solution for your tastes perhaps, but IMO any browser without javascript is totally useless.

Quote
sure java script is miss, and OWB is for MOS OS4 AROS users great, because better wait 5 sec instead of cant show page.but when i see that the page is show on 2 sec(the page layout is done and can scroll) on my windows box with all browsers, i think its not usefull to spend lots work in OWB before the speed get same as other browsers.
I don't know what you're talking about. At least MorphOS OWB is instant here.

Quote
I like too have Java script and problem of later show on OWB have nothing to do with Java script.I deactivate it for test in AROS OWB, but get not faster.
This argument is null and void. Netsurf still lacks javascript.

And I still don't see any performance problems in OWB compared to Netsurf. You must've been benchmarking wrong things again.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: x303 on December 20, 2009, 12:42:39 PM
Quote from: bernd_afa;534557
and is also not listet on wikipedia, do not motivate me for OWB.
Eh, here: http://en.wikipedia.org/wiki/Origyn_Web_Browser


It would be nice to have owb simply because you can't live without javascript anymore. And we can't wait a few years on it to appear on netsurf.


x303 :D :D :D
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 20, 2009, 04:08:59 PM
Quote from: bernd_afa;534560

...
and BTW, somebody who is intresting on Port to 68k,  have fab and me ask about OWB, he install now amidevcpp and want it do in cygwin.i send him my full includes compilers with all libs i have.
...


That was me! :)
I gave up because my amiga programming knowledge is too limited.

Quote

If somebody want port it, i help.
but i do no work, when nobody want do something.he can also try himself to compile Cairo.


If u compile the libraries, maybe someone else is awaiting just this; open-source is NOT a one-man job (at least in linux world :) )
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Crumb on December 20, 2009, 04:56:26 PM
@thread

IMHO it would be better to start from Fab's MorphOS sources as he has done a wonderful port. You'll probably have to deal with some MUI4 methods and specific MorphOS functions but I think it's better than trying to update Joerg's port.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 20, 2009, 06:44:47 PM
>Huh? So you seriously think just because there's software X in linux distros it somehow >measures popularity of that said software?

there are also some patches for windows/linux and OWB old versions, but nobody do actual OWB release for that system.

if so, i can maybe install a small linux on vmware and test OWB here in compare with Linux netsurf.

and btw, can you tell me page times, and links where OWB is faster ?
I see the video from Fab, here the load time of full page was good, but the time the page was show first and can scroll, was much longer as on netsurf and other browsers.this cant be problem of internet, because in this early stage very few data is need from Internet.

you should not test the old Netsurf for MOS on reuters or some other page.because this old Versions have Bug that redraw is done too often when there is a image skip delay > 1 sec.thats fix in newer versions 68k have but MOS not.

>If u compile the libraries, maybe someone else is awaiting just this; open-source is NOT a >one-man job (at least in linux world  )

>I gave up because my amiga programming knowledge is too limited.

amiga programming knowledge is mostly not need.you need Unix ccmake and C++ knowledge and much patience.

>The reasons for AROS OWB slowness were explained already.

You mean because of SDL ?

a browser is simple written, it render the page in ram with CPU commands.so the time the page is show first is same.maybe the transfer to GFX Card is in SDL only 4 fps but still here 4 fps are 0,25 sec, and we talk about several seconds that OWB show the page later as other browser.

also i see OWB on AROS show same fast as netsurf the whole page.but netsurf or other browser show 2-3* sooner the page.so i think its OWB dependent.

I dont know how can MOS version be faster.

using webkit is ok, but the trick is use it in best way asynchron and here it seem the OWBAL have some Problems.

So wy not accept this and report that Problem in OWB ML to discuss instead say OWB is fast and great ?

the result can only be a better OWB and it get more attractive and then come upto date Linux builds maybe.That OWB have no diskcache all browser have is also a Limit.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 20, 2009, 07:30:07 PM
Quote from: bernd_afa;534601
>Huh? So you seriously think just because there's software X in linux distros it somehow >measures popularity of that said software?

there are also some patches for windows/linux and OWB old versions, but nobody do actual OWB release for that system.

if so, i can maybe install a small linux on vmware and test OWB here in compare with Linux netsurf.

and btw, can you tell me page times, and links where OWB is faster ?
I see the video from Fab, here the load time of full page was good, but the time the page was show first and can scroll, was much longer as on netsurf and other browsers.this cant be problem of internet, because in this early stage very few data is need from Internet.

you should not test the old Netsurf for MOS on reuters or some other page.because this old Versions have Bug that redraw is done too often when there is a image skip delay > 1 sec.thats fix in newer versions 68k have but MOS not.



My video was actually quite slow comparing to what it can achieve on a proper link (or proper router). My router often gets unresponsive after a few days uptime. Yet, it's still much faster than what i could get from netsurf (both MorphOS and current 68k version).

Quote

>The reasons for AROS OWB slowness were explained already.

You mean because of SDL ?

a browser is simple written, it render the page in ram with CPU commands.so the time the page is show first is same.maybe the transfer to GFX Card is in SDL only 4 fps but still here 4 fps are 0,25 sec, and we talk about several seconds that OWB show the page later as other browser.

also i see OWB on AROS show same fast as netsurf the whole page.but netsurf or other browser show 2-3* sooner the page.so i think its OWB dependent.

I dont know how can MOS version be faster.


First, the plain SDL version is much slower at scrolling for the reason given earlier and then, regarding network transfer, it's mostly down to Curl usage (the default in OWB is not optimal for fast connections) and underlying TCP/IP stack too. In the MorphOS port, i also moved the whole network part to a dedicated thread to avoid blocking UI during DNS resolution or other network transfers, which effectively gives a better feeling too.

Quote

using webkit is ok, but the trick is use it in best way asynchron and here it seem the OWBAL have some Problems.

So wy not accept this and report that Problem in OWB ML to discuss instead say OWB is fast and great ?

the result can only be a better OWB and it get more attractive and then come upto date Linux builds maybe.That OWB have no diskcache all browser have is also a Limit.


I don't see why the availability of linux builds has anything to do with the quality of a program. And for your information, OWB is mostly an API layer above WebKit designed to port and use it more easily. There's no major functional difference with plain WebKit (which is used in Safari, Chrome and other browsers). And about diskcache, well, it's not a big these days, and there's a memory cache, so it's ok during a same session, anyway.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 20, 2009, 07:51:49 PM
Bernd, have u heard Artur lately ? I tried to contact him, but didn't get an anwser.
He may have a go at porting Fab's version. Or he's secretly doing it already ?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 20, 2009, 07:53:39 PM
Quote from: bernd_afa;534601

can you tell me page times, and links where OWB is faster ?

OWB is faster everywhere. And, yes, you can scroll down instantly.

Your benchmark methods are flawed I guess. You really ought to rule out internet connectivity issues from your tests. Free hint: Use local copies of web pages you test.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: amigadave on December 20, 2009, 11:38:54 PM
Quote from: alexh;534490
If you want to access the web from your Amiga and you have another PC on the network (Mac, Linux, Windows etc.) there are tools that will let a networked Amiga open a web-browser on those machines and display it on the Amiga screen.

With the Amiga hard drives mounted via SAMBA on the host machine (so you can save directly to the Amiga) it looks and feels almost like a native browser only much faster and upto date.

Just google for Amiga & VNC or Remote Desktop.

I have a HP NAS drive (EX470) running Windows Server 2003 which is always on and I click on an icon on the Amiga and a remote FireFox 3.5.6 window opens. Because it is remote everything just works like it would on a PC, Flash, Shockwave, Java etc.

This is going to be one of my next projects.  Is a graphics card required in the Amiga to run VNC or Remote Desktop?  It would be nice if it was not and was usable on my A1200 w/IndivisionAGA display.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 09:14:34 AM
>Your benchmark methods are flawed I guess. You really ought to rule out internet >connectivity issues from your tests. Free hint: Use local copies of web pages you test.

I test too with cache proxy.but i get that not faster.
But your idea is good, maybe there is somewhere a benchmark HTML Page that can run from Harddrive and we can use for profiling browsers ?

I still dont understand wy OWB on 68k/ AROS need so much time to show a page without images.he need only load the main file parse and render it.

>Yet, it's still much faster than what i could get from netsurf (both MorphOS and current >68k version).

Is there on MOS a CPU throttle Software ?.I guess when you throttle your CPU that it is only 100 MHZ fast, then you notice too that OWB is lots slower.

When caches are small no 2. Level cache is here it slowdown more.
Maybe somebody can show a pageload on a efika at http://www.reuters.com or a file of that ?

I think with such a system you notice more the diffrence between OWB and netsurf.

I never see a fast C++ Program.dont forget C++ need lots of Overhead often copy data and this need larger caches.also using cairo increase cache misses.

and thats the reason that the CPU developers notice that a Cache > 1 Megabyte still give some speedup-

look at memory usage of OWB, thats another problem.it use lots mem.much more as netsurf.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 21, 2009, 09:44:04 AM
Quote from: bernd_afa;534689

...
I guess when you throttle your CPU that it is only 100 MHZ fast, then you notice too that OWB is lots slower.
...


Sorry, Bernd, Owb is not meant for real A4000 machines; owb 68k target are solutions like this one: http://xamiga.net/ on a modern PC.

Plus, I don't understand why u shouldn't believe Fab when he says that his MOS version is faster than AROS one ?
I renew the proposal: if u don't want to do all the job yourself, compile just the libraries, then Fab (if he wants) or others interested can cross-compile owb.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: wawrzon on December 21, 2009, 10:56:05 AM
as i said many times already os4 version of owb doesnt usually feel slower than os4 port of netsurf. i believe the opposite is the case. but then bernd has some principal views that one is unable to discuss with. one of them is c++ apps are slow, another oop is crap, and third owb is slower than netsurf, just to name a few. anyway being practically illiterate on the subject im not in a position to prove it wrong, can only voice an opinion based on practical experience.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 11:15:19 AM
>Plus, I don't understand why u shouldn't believe Fab when he says that his MOS version >is faster than AROS one ?

I see this video from Fab when load http://www.cnn.com and with AROS OWB i get same times(7-8 sec) to show page first.safari/chrome is lots faster around 2-3 sec.netsurf on my winuae need 4-5 sec.not optimal, but better as OWB on AROS, netsurf Team is working to add a cache.I think when its done, netsurf is speed up.

http://fabportnawak.free.fr/vids/cnn.mpg

sure he tell situation is not optimal, but normaly i think when doing a vídeo before first try once and then do it again to see if time is same and reproducable.and when his OWB really is so fast as firefox or other windows webkit browsers, then its clear notice if a browser need 2-3 sec to show a page or 7 sec.

Also you can see on Video, he do no reload.he type the name in, its possible that there is something on cache.

>Sorry, Bernd, Owb is not meant for real A4000 machines; owb 68k target are solutions >like this one: http://xamiga.net/ on a modern PC.

but still, it run slower because of slow internet,OWB load too much data from internet before he begin to show the page.also when press the page back button the load of the last page is same slow as a full load.this i see on 68k OWB and AROS OWB.

normaly when press page back the page is show in around 0,5 sec complete.On netsurf some pages are show fast some not, but as netsurf Team told teh cache is currently broken and is rewritten to do disk caching too.

>compile just the libraries, then Fab (if he wants) or others interested can cross-compile >owb.

Have you ask zero hero if he compile Cairo ?.Here are lots of libs for 68k

http://megacz.back2roots.org/portsbttr.html

I currently have more fun to enhance stormwizard to have same nice GUI look as MUI/zune programs.I get ok from Haage&Partner that i am allowed to enhance it and latest source from Alinea on 12.09 to release the lib and the GUI Editor for free.

I do the enhance with frameiclass so all OS or GUI Patches that support this frame types can get nicer system conform look and it depend on the prefs of the system and wizard library need no Skin preferences.

AFA use zune/MUI render features for frameiclass

Only not good Limit i must accept, the wizard source can not go opensource and i can not give it to other people and its currently not allow to make a AROS or MOS native Version.but lets see what the future bring, when the red versus blue war end, maybe the Limits change.

#define FRAME_DEFAULT      0
#define FRAME_BUTTON      1
#define FRAME_RIDGE      2
#define FRAME_ICONDROPBOX   3
4 is the prop container and 5 prob knob.OS4 support that too(See sdk)
what stormwizard GUI System is you can see on aminet(it currently not work) when you search for storm Wizard.

Here is a screenshot how Arteffect look on AFA with new wizard.library

http://www.daten-transport.de/?id=tBqry9fq5tF4
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 21, 2009, 11:24:52 AM
Quote from: bernd_afa;534696
...

Have you ask zero hero if he compile Cairo ?.


What's the problem with Cairo ?
Or find someone else to compile it. Do u know other 68k devs, besides Artur, who can do it?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 11:58:52 AM
>as i said many times already os4 version of owb doesnt usually feel slower than os4 port >of netsurf. i believe the opposite is the case.

here you should bring time values.maybe OS4 netsurf is slower as 68k netsurf.

when use cnn page

after what time netsurf/ OWB on OS4 show the page first ?
and then the time netsurf 68k do.best is test 3 times

You have a PC, maybe you can compare AROS OWB and OS4 OWB.
OS4 OWB is not so enhanced as MOS OWB

But when MOS OWB is faster as OS4 OWB, then there should some values here.there was some OS4 MOS compare benchmarks, but no OWB compare benchmarks.

OWB on both system is able to render page from file, or maybe use a cache proxy.

sure i know best is when i test that on a MOS system, but i really dont want buy a MOS able System to see that there are no wonders possible

>What's the problem with Cairo ?

I dont know, i have not compile it.there is a AROS version from 2006 nobody use.and now when somebody really need cairo, he need the newest version, maybe here is change something.Or where is the source for MOS Cairo

Cairo is a backend and i think it need some platform dependent graphic stuff.

But i remember when firefox change to Cairo render it get too slower.I think best is use SDL and fix teh scroll problem by using HWsurface and sdl blit command.

maybe that give a speedboost and fewer memory usage on efika
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 12:32:46 PM
>Or find someone else to compile it. Do u know other 68k devs, besides Artur, who can do >it?

as i told zero_hero can do it,
also amistuff can do it, diegocr and many i forget.

i think on 68k there are more devs that can do it.But there seem no perfect browser for amiga currently here.

the browser thats done on MOS or OS4 are done from the OS Developers.On a commercial  OS its important to have a better than nothing solution as fast as possible to sell more systems.

but developers that develop for fun, want do something better usefull or when they get money(see AROS Port)

And as can see, for 68k there are no bounties or something else as on AROS or OS4 or MOS.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 21, 2009, 12:44:32 PM
Quote from: bernd_afa;534703
>Or find someone else to compile it. Do u know other 68k devs, besides Artur, who can do >it?

as i told zero_hero can do it,
also amistuff can do it, diegocr and many i forget.


Ask one of them (or all of them) to co-operate in the porting; more hands are better than one.

Quote

  But when MOS OWB is faster as OS4 OWB
...


Fab already said MOS OWB is faster. I say: let's trust his words.


Quote


the browser thats done on MOS or OS4 are done from the OS Developers.On a commercial  OS its important to have a better than nothing solution as fast as possible to sell more systems.

but developers that develop for fun, want do something better usefull or when they get money(see AROS Port)

And as can see, for 68k there are no bounties or something else as on AROS or OS4 or MOS.


Fab asked no money in order to do it.
Do u want to be paid ? Is this the "problem" ?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 02:20:04 PM
>more hands are better than >one.

yes thats right and so i think there should more working together to bring a perfect browser.

if the MOS OWB is really faster, then maybe somebody add the enhancements to AROS or add it in the Main OWB Core.

and when i then see it work same fast as other browsers, then i am more motivate to compile it for 68k.

I need not hurry up, and can wait and look whats happen in 1 year if it is then more easy possible to port OWB.

>Fab already said MOS OWB is faster. I say: let's trust his words.

Fab is a Morph OS Developer.
When you ask the Microsoft Browser Developers, i think he answer too that Internet explorer is fastest ;-)

And same is with the Apple Safari Developers i think. ;-)

But i can test IE and Safari and notice that Safari is faster.

>Fab asked no money in order to do it.
Yes, but we not know if he get money from MOS sells.

that he accept no bounty can also better Marketing.

Or what do you think when Apple/MS make a browser that miss since some years the importants things as a download progress bar and more.
and there is on Apple/MS Homepage a button for bounty ?

>Do u want to be paid ? Is this the "problem" ?

No, i dont want paid, i want fun on coding and something i can use myself.netsurf i can use on some pages currently, but AROS OWB is too slow to use for me.I test on vmware and Icaros.

I like more the AROS Version because it use no Cairo and is a MUI class.So browser can use with every application.and because its a class there is possible to build a amiblitz GUI for example.

because i use winuae i can use windows browsers until the browser situation in amiga land look better.maybe MOS OWB get as MUI class some day.then its more easy portable

maybe somebody add the libcss/libhtml of netsurf to aweb or so.

so before spend lots work, better wait.netsurf is here and because its written not in C++ it can be more easy get faster, it can be much faster compile etc.

I have program a peephole Optimizer for amiblitz and i know also what much additional code the CPU on C++ must execute, to run this nice OO code.

Many developers that use C++ dont know anything abput how a CPU work, they never start a profiler or have any glue how fast a 1 GHZ CPU can be.

for example.with a 50 MHZ 68060 you can compile perfect paint complete (source have 60000 lines in amiblitz in 12 sec.then you can start it and after load 4 sec later it show the main window.

so i really dont understand wy a page that contain some text and parse some 200 lines of html code on a classic need over 18 sec show.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 21, 2009, 02:29:25 PM
Quote from: bernd_afa;534696
>Plus, I don't understand why u shouldn't believe Fab when he says that his MOS version >is faster than AROS one ?

I see this video from Fab when load http://www.cnn.com and with AROS OWB i get same times(7-8 sec) to show page first.safari/chrome is lots faster around 2-3 sec.netsurf on my winuae need 4-5 sec.not optimal, but better as OWB on AROS, netsurf Team is working to add a cache.I think when its done, netsurf is speed up.

http://fabportnawak.free.fr/vids/cnn.mpg

sure he tell situation is not optimal, but normaly i think when doing a vídeo before first try once and then do it again to see if time is same and reproducable.and when his OWB really is so fast as firefox or other windows webkit browsers, then its clear notice if a browser need 2-3 sec to show a page or 7 sec.

Also you can see on Video, he do no reload.he type the name in, its possible that there is something on cache.



When my router is responsive, cnn displays in 3-4s. But should we remind you once again that network speed and latency is a *very* important factor in this kind of benchmark?

And no, there was no cache involved in my test.


Quote

but still, it run slower because of slow internet,OWB load too much data from internet before he begin to show the page.also when press the page back button the load of the last page is same slow as a full load.this i see on 68k OWB and AROS OWB.

normaly when press page back the page is show in around 0,5 sec complete.On netsurf some pages are show fast some not, but as netsurf Team told teh cache is currently broken and is rewritten to do disk caching too.


There's a page cache in WebKit, that just happens to be disabled by default in OWB (understandable considering their embedded orientation). When page cache is enabled, it obviously displays in a fraction of second, if the page is still in cache.

Quote

I dont know, i have not compile it.there is a AROS version from 2006 nobody use.and now when somebody really need cairo, he need the newest version, maybe here is change something.Or where is the source for MOS Cairo

Cairo is a backend and i think it need some platform dependent graphic stuff.

But i remember when firefox change to Cairo render it get too slower.I think best is use SDL and fix teh scroll problem by using HWsurface and sdl blit command.


Cairo/Pixman straight-compiles (unless you add hw acceleration of course, then it's a lot more work). And there's a reason to use Cairo over SDL too. In OWB, they also moved from the SDLgfx graphics backend to Cairo one, because WebKit just needs a capable gfx library and SDLgfx isn't up to the task. So in current OWB SDL implementation, rendering is done with Cairo and the result is blitted with SDL.

Quote

i think on 68k there are more devs that can do it.But there seem no perfect browser for amiga currently here.

the browser thats done on MOS or OS4 are done from the OS Developers.On a commercial OS its important to have a better than nothing solution as fast as possible to sell more systems.

but developers that develop for fun, want do something better usefull or when they get money(see AROS Port)

And as can see, for 68k there are no bounties or something else as on AROS or OS4 or MOS.


I'm not sure what you imply there, but should I remind you I didn't charge nor ask anything for OWB? It wasn't a bounty either.

Also, to me, it seems OWB+SDL and NetSurf+SDL ports are closer to the "better than nothing" quick hacks than the other NetSurf or OWB ports on OS4/MorphOS. At least, there was some real work involved in the latters to build a GUI around the engine, unlike SDL ports.

@0amigan0
Quote

Fab already said MOS OWB is faster. I say: let's trust his words.


Actually, I didn't say MorphOS OWB was faster than OS4 OWB. They're roughly the same I'd say, both faster at scrolling than AROS or Linux ports (which is normal, since they lack a proper scroll method). The main difference would be MorphOS port uses a thread for network, which avoids blocking UI.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 21, 2009, 02:35:35 PM
Quote from: bernd_afa;534715
...

because i use winuae i can use windows browsers until the browser situation in amiga land look better.maybe MOS OWB get as MUI class some day.then its more easy portable

MOS OWB is *already* a MUI application (AFAIK).
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 21, 2009, 02:45:18 PM
Quote from: bernd_afa;534715
>more hands are better than >one.
>Fab asked no money in order to do it.
Yes, but we not know if he get money from MOS sells.

that he accept no bounty can also better Marketing.

Or what do you think when Apple/MS make a browser that miss since some years the importants things as a download progress bar and more.
and there is on Apple/MS Homepage a button for bounty ?


I don't request anything from MorphOS sales, and if I did, it would certainly not be about OWB. I'm generally not at ease with requesting money for an opensource software where i only wrote 0.1% of the whole code.

Maybe you're thinking about that other open(soon-to-be-closed)-source bounty-based browser ported by some OS core developers. :)

Also, I'm not sure what you are referring to about missing download progress, but i hope you don't refer to my port.

Quote

I like more the AROS Version because it use no Cairo and is a MUI class.So browser can use with every application.and because its a class there is possible to build a amiblitz GUI for example.

because i use winuae i can use windows browsers until the browser situation in amiga land look better.maybe MOS OWB get as MUI class some day.then its more easy portable

Fine with that. It would probably be easier to port since you wouldn't have to adapt some MUI4 calls. You'd lose all the other features that are not available in AROS version, though.
About the class, it's certainly something nice to have, but I think the current AROS class crashes if there's more than one opener, which is a bit limiting. :)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 03:40:24 PM
>Also, I'm not sure what you are referring to about missing download progress, but i hope >you don't refer to my port.

I mean not your Port, i know MOS and AROS OWB have a download progress bar.

>MOS OWB is *already* a MUI application (AFAIK).

I mean that OWB is a class, simular to the MUI HTML class, but of course can show all.
Then its possible to add code easy that simple Mail and YAM can show Mails in HTML too.

but because simple Mail and Yam is a multiplatform AOS Program that run on 68k MOS OS4 AROS, it does not help when in amigaland no working together is possible and every side do their own browser.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 21, 2009, 03:51:18 PM
>About the class, it's certainly something nice to have, but I think the current AROS class >crashes if there's more than one opener, which is a bit limiting.  

the bounty rule is say it should work.if you know a problem in design, you should report it.when it not work, its also not possible when run 2 or more OWB at same time.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 21, 2009, 04:26:59 PM
Quote from: bernd_afa;534729
...
but because simple Mail and Yam is a multiplatform AOS Program that run on 68k MOS OS4 AROS, it does not help when in amigaland no working together is possible and every side do their own browser.


OK, we all know this. Now, can u try, at least try to straight-compile the libraries required by Mos Owb ? Yes or No ?

If No, well... we can close here this thread.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 23, 2009, 08:44:38 AM
Quote from: Fab;534718


Actually, I didn't say MorphOS OWB was faster than OS4 OWB. They're roughly the same .


If you see no diffrence, then MOS OWB have same Problem.Here can too read from 2 OS4 dev that OWB load much of the page.try also the link on MOS OWB

other browsers need not so much load before a page is usable(mean scrollable/movable)

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30292&forum=32#526721

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30292&forum=32#526765

this page reload in 1-2 sec on windows browsers.need test how long it take on AROS.

>When page cache is enabled, it obviously displays in a fraction of second, if the page is >still in cache.

And how can enable this, so i can check in AROS/68k sources ?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 23, 2009, 08:59:26 AM
@bernd_afa

You sound like a broken record.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: utri007 on December 23, 2009, 10:49:33 AM
You are allways so polite and friendly.

[edit by admin: personal attack.]
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 23, 2009, 11:11:59 AM
@utri007

if you have a problem with something I said, please specify what it was.

otherwise stfu. and keep our relatives out of it.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Argo on December 23, 2009, 12:32:32 PM
Okay, let's leave people's mothers out of this. Not personal attacks either.
Keep it civil and constructive criticizim.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 23, 2009, 12:49:08 PM
>You sound like a broken record.

i must post only facts because told MOS OWB is fast as other browsers.

It does not help to make things better or attract more Users to Amiga OS NG Systems when all is tell great, and features that are miss, is tell they are not need.there arte lots other sytems out that offer this features.and thats maybe the main Reason wy AOS have since all this years not at least 5000 Users from the several Million Users Amiga have in the past.

such a strategie can only work in a Monopol World, but Amiga have no Monopol.

maybe when all work together OWB can get equal in compare to other browser, and all users save time when then surf
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 23, 2009, 02:16:25 PM
Quote from: bernd_afa;534886
If you see no diffrence, then MOS OWB have same Problem.Here can too read from 2 OS4 dev that OWB load much of the page.try also the link on MOS OWB

other browsers need not so much load before a page is usable(mean scrollable/movable)

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30292&forum=32#526721

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30292&forum=32#526765

this page reload in 1-2 sec on windows browsers.need test how long it take on AROS.

There's a difference between speed and responsivity, but that must be too subtle for you, and these 2 OS4 dev have no idea about the real issue. It's not about loading too much data, really. You might consider WebKit is the very same engine as chrome and safari use, and OWB didn't change its behaviour there. The problem is in the network handling, like i said several times before.

In MorphOS port, I moved network code to a dedicated thread, which avoids the responsivity issues listed in that thread. So i can act on the document as soon as the document(+css?) is received, just like it should be. I can even profile it precisely with WebInspector, actually.

About your (meaningless) benchmark, that page reloads in 2 seconds on OWB MorphOS, so what?

Quote
>When page cache is enabled, it obviously displays in a fraction of second, if the page is >still in cache.

And how can enable this, so i can check in AROS/68k sources ?

In webcore settings, the name is very explicit (pagecache and cachemodel).
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 23, 2009, 06:55:28 PM
>There's a difference between speed and responsivity, but that must be too subtle for you,

I know that, but when you read th second post, he write that OWB on OS4 is too responsive.

""""
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=30292&forum=32#526731

I have to say that isn't true. I can click on links, scroll, etc while it is still loading some images (and maybe other stuff). But it does seem to need to load a lot of the page first, before rendering
"""

As soon OWB on 68k or AROS show the page the first time it is responsive, at least with a delay of 200 ms.but until OWB show a page it need much more time than other browsers.

when i try 5 times to reload cnn on safari and get times to first show of 2-3 sec and then try with OWB and  AROS and then i need 7-8sec then its a huge noticable diffrence.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 29, 2009, 11:32:42 AM
@Fab:

You have a Linux cross-compile environment ready to go.
YOU might try your hand at it, if Bernd doesn't want to hear.

You surely will be idolized by us 68k freaks. :)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on December 29, 2009, 03:34:51 PM
Quote from: 0amigan0;535334
@Fab:

You have a Linux cross-compile environment ready to go.
YOU might try your hand at it

You forget that the target for that cross compiler is MorphOS PPC.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 29, 2009, 04:27:28 PM
Quote from: Piru;535353
You forget that the target for that cross compiler is MorphOS PPC.


I didn't know that.
well, never mind. :)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 29, 2009, 04:46:57 PM
Quote from: 0amigan0;535361
I didn't know that.
well, never mind. :)

Indeed, my setup isn't ready at all for a 68k target. Not that it would be impossible of course, but sdk stuff always a tedious task (at least for me :)).

In any case, if a 3.x developer with a "correct" gcc4 + libnix setup (no ixemul, thanks :)) wants to give it a try, i'd be glad to help him. It won't be a very trivial task, so it needs a bit knowledge (i.e more than configure/make :)).

Some things would have to be adapted or "degraded" to work on OS3.x/MUI3.8 (network thread support, MUI tabs or list methods, ...), but it's just some additional work, nothing impossible (and if it's too scary, one could always start with the AROS port first, which should compile more easily on 3.x). About the speed it would have on a real amiga, i couldn't tell though, but it would certainly be substancially than current SDL version in any case, at least regarding scrolling.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on December 30, 2009, 12:42:00 PM
@Fab:

I've found this cross-compiler setup guide http://cross.zerohero.se/os3.html
It mentions gcc 3.4.0; where can I found gcc 4 ? Are the tools he use, up-to-date ??

Can I use these mui 3.8 developer files http://aminet.net/dev/mui/mui38dev.lha ? Provided of course I can manage to setup the cross-environment. :)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on December 31, 2009, 12:25:55 PM
>but it would certainly be substancially than current SDL version in any case, at least >regarding scrolling.

I look at the OWb 68k 1.2 (SDL version) and test without JIT.Scrolling is very slow need ca 1,3 sec for 1 frame.netsurf scroll also when pixel format not match with at least 8 fps on my System and disable JIT.

But when use the JIT OWB 1.2 scroll with 10-15 fps and thats lots faster as AROS Version run on my System.AROS Version scroll only with 3 fps on vmware.

Now when use OWB 68k 1.4 that use a enhanced SDL Version.It scroll on my system without JIT with 10-15 fps thats fast as netsurf do when pixelformat in SDL is ok.

But!!!

The OWB 1.4 need for a page load 2* longer as the OWB 1.2 full SDL Version.

So i think best is to fix in the SDL Version the pixel Match Problem.It seem SDL is in general very slow when it must convert pixel Formats.Same Problem is maybe with Cairo.

on 68k some GFX Cards work in RGBA some in BGRA.Cairo use same as opengl intern RGBA Pixel Format.any other must convert and get speedloss.

Do you know if its possible to set the Pixelformat of the OWB render engine to any Pixelformat, or use the OWB render engine a fixed Format ?

the Netsurf Render engine work only in the ARGB Pixelformat.This Pixelformat does Cairo not support native, so Cairo get some speedloss because it need convert internaly

What Pixel Format MOS use ?
Can MOS open screens of diffrent Pixelformats ?
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on December 31, 2009, 02:28:14 PM
Quote from: bernd_afa;535535
>but it would certainly be substancially than current SDL version in any case, at least >regarding scrolling.

I look at the OWb 68k 1.2 (SDL version) and test without JIT.Scrolling is very slow need ca 1,3 sec for 1 frame.netsurf scroll also when pixel format not match with at least 8 fps on my System and disable JIT.

But when use the JIT OWB 1.2 scroll with 10-15 fps and thats lots faster as AROS Version run on my System.AROS Version scroll only with 3 fps on vmware.

Now when use OWB 68k 1.4 that use a enhanced SDL Version.It scroll on my system without JIT with 10-15 fps thats fast as netsurf do when pixelformat in SDL is ok.



The thing is when you scroll 20 pixels or so per step, it should be much faster, when a backstore scrolling method is used, which isn't implemented in the SDL backend.
For instance, here, on peg2, on average, with a 1024x768 window, rendering+scrolling a 1024x1 line takes about 2 ms. Rendering the whole page can take about 40-100ms (depending on content). So, implementing a proper scrolling method can effectively make scrolling dozens times faster (of course, when you scroll with a very big step (page per page or more), it doesn't have any benefit anymore).

Quote

But!!!

The OWB 1.4 need for a page load 2* longer as the OWB 1.2 full SDL Version.



No idea, but like i said several times, there are many things to be done in network layer.


Quote

So i think best is to fix in the SDL Version the pixel Match Problem.It seem SDL is in general very slow when it must convert pixel Formats.Same Problem is maybe with Cairo.

on 68k some GFX Cards work in RGBA some in BGRA.Cairo use same as opengl intern RGBA Pixel Format.any other must convert and get speedloss.

Do you know if its possible to set the Pixelformat of the OWB render engine to any Pixelformat, or use the OWB render engine a fixed Format ?

the Netsurf Render engine work only in the ARGB Pixelformat.This Pixelformat does Cairo not support native, so Cairo get some speedloss because it need convert internaly

What Pixel Format MOS use ?
Can MOS open screens of diffrent Pixelformats ?


The SDL version of OWB didn't use cairo at all. It used sdlgfx to render (since a few weeks, it can also use cairo to render, blitting still happens in SDL of course).

And by the way, Cairo supports ARGB pixel format.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on January 02, 2010, 10:16:56 AM
>when a backstore scrolling method is used, which isn't implemented in the SDL backend.

but its easy to do so.
see code in netsurf

http://source.netsurf-browser.org/trunk/libnsfb/src/frontend_sdl.c?r1=9719&r2=9720

and when use HW surface the System blitter is used

> No idea, but like i said several times, there are many things to be done in network layer.

when a test is done 10 times and always 10 times 1.2 work 2* faster as 1.4 and times do not differ more than 10-20% each test then can in real world 100% sure 1.2 is faster.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: 0amigan0 on January 10, 2010, 11:35:26 AM
@Fab:

Can I use this guide ( http://www.zerohero.se/cross/os3.html ) in order to setup a cross-compiler environment ??
Are the tools indicated, up-to-date ??
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: samo79 on January 11, 2010, 06:07:57 AM
Quote from: bernd_afa;534553
If OWB is faster, then wy netsurf is a google summer of Code projecrt since 2 years and there is no OWB Version in a Linux distri ?.the big disadvantage of netsurf is currently missung java script but its in Linux Distri.

netsurf is in Linux distri.Wy not OWB ?


Because OWB is a Sand-Labs project initially dedicated only to CE device and it was/is builded around WebKit, you didn't found it on Linux yet just because they are use a more complete core or complete browser around the same WebKit core

Having said that you can't compare it with NetSurf that it is actually a "spare time" project maded by a little and passionate community, mostly RiscOS developers ...

As Fab say actually OWB on OS4/MOS are MUCH faster than any NetSurf release and it is constantly updated.

What I like to see (maybe one day) is a MUI version that can cover all the Amiga like systems, expecially for AmigaOS 3.x, that because when OS4 Reaction class are constantly updated, 68k class not :-)

P.S.
I'm a NetSurf 3.x betatester on OS4 so I know very well what I say ;-)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on January 12, 2010, 06:28:51 PM
Quote from: samo79;537627

P.S.
I'm a NetSurf 3.x betatester on OS4 so I know very well what I say ;-)


You cant compare the OS4 version that use another layer (Cairo) with the 68k Verison.In some post is written that 68k netsurf was faster as OS4 netsurf btw.

netsurf 68k can also show files offline when do file://dh2:test.html to show dh2:test.html.
but must be sure that all references gone to file

I have measurement enough and i notice always the same that other have written too (link in this thread).OWB show a page much later as all other browsers.I see this on a Video too.If this video is not good, then wy there is no better.

And its near impossible that this in MOS or OS4 is better

In Amiga land its known everybody want have the best system and you cant believe what cant see with own eyes.

But i not buy a MOS or OS4 system to see same as i can see on AROS.Maybe OWB scroll on MOS or OS4 faster, but how many time is need to show the page first doesnt depend on GFX Speed because also the AROS Version can render a page in 300 ms.

>What I like to see (maybe one day) is a MUI version that can cover all the Amiga like >systems, expecially for AmigaOS 3.x, that because when OS4 Reaction class are >constantly updated, 68k class not :-)

I like to see this too, but as long every system want fight for more user with a better browser its lots work when all systems do their own Version.

And i have hope in 1-2 years when the MOS/OS4 developers maybe see there cant make enough money with the OS they stop the lots work spend on browser.

and then the last existing users/dev maybe do together bring a actual browser.
I can also port in 4 years OWB to 68k, maybe then OWB is better
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Fab on January 12, 2010, 07:33:18 PM
Quote from: bernd_afa;537896


I have measurement enough and i notice always the same that other have written too (link in this thread).OWB show a page much later as all other browsers.I see this on a Video too.If this video is not good, then wy there is no better.

And its near impossible that this in MOS or OS4 is better

In Amiga land its known everybody want have the best system and you cant believe what cant see with own eyes.



Will you finally understand that there are some differences between all these OWB ports? In MorphOS port, I changed a few things because the default values were just not adapted for "modern" inet speeds (that small change can actually make transfers 20 times faster, going from 400kB/s to 8MB/s locally ...). And more importantly, I threaded the network to give a better responsivity.

Recently, the AROS port implemented backstore scrolling, and also the trivial (but needed) network speedup. So you might try it again, and finally understand that the plain SDL OWB port doesn't mean nothing at all, performance-wise, and that some work is needed, just like in netsurf...

Quote

I like to see this too, but as long every system want fight for more user with a better browser its lots work when all systems do their own Version.

And i have hope in 1-2 years when the MOS/OS4 developers maybe see there cant make enough money with the OS they stop the lots work spend on browser.

and then the last existing users/dev maybe do together bring a actual browser.
I can also port in 4 years OWB to 68k, maybe then OWB is better


Or maybe that in 4 years, you'll have finally understood that not everyone wants to develop on a dead OS, and that if you really want it, you could port it yourself, instead of waiting for someone to do it. And by the way, I offered my help for a 68k port, and someone contacted me about it.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Piru on January 12, 2010, 08:01:52 PM
Quote from: bernd_afa;537896
And i have hope in 1-2 years when the MOS/OS4 developers maybe see there cant make enough money with the OS they stop the lots work spend on browser.

As "making money" isn't the goal I'm afraid we can't grant that wish.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: samo79 on January 12, 2010, 08:10:18 PM
Pardon but how you can't make a speed comparision in a YouTube video ??

I tried also NetSurf on 68k but I can't say if speed are really better than the OS4 release as I play with it on an old x86 PC in emulation via WinUAE, but aniway in one thing i'm sure 100%, OWB under OS4 is perfectly for navigation and there aren't any issue on this way

Maybe we can speak about UI problems or other issue but not about time rendering

Latest OWB's rendering on my Flex 800 is faster than Firefox 3.0 on an high-end PCs, but you are free to believe anythings ;-)
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: kolla on January 12, 2010, 10:37:38 PM
Quote from: Piru;537908
As "making money" isn't the goal I'm afraid we can't grant that wish.


So the price tag of MorphOS is there to keep the number of whining users down - that's what I've been suspecting :lol:
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: Karlos on January 12, 2010, 11:03:52 PM
No need to knock C++ in this thread. It's perfectly possible to write fast C++ code. It's my preferred development language, even on m68k amigaos.

One of these days I'm going to have to write all the strange things I discovered in gcc 2.95.3 on m68k :) Some of them are quite useful.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: unusedunused on January 13, 2010, 10:17:52 AM
>As "making money" isn't the goal I'm afraid we can't grant that wish.

What is the goal then, when MOS is not opensource or at least, some Part as libnix or MUI4 can port to other AOS to make it more easy to write Programs that run on all Amiga Systems ?

There are lots devs that support all systems with same program and common libs help alot to use modern features.

So when all side (MOS OS4) release their closed source stuff, the few existing AOS devs can add more features in shorter time and the programs look better in compare to Linux or windows Software.

also there need not several devs that work on the same stuff.so a dev can do a other program and so more programs come.

but what currently happen in amigaland is this.

A ship sink and there are 4 guys left, but instead they join 1 boat, every guy use his own boat.

but every guy must sleep, so they move slower forward.

But in amigaland its of course more complicate because every "boat" have his advantages but also lots disadvantages that avoid to join the boat.So its impossible to get all in 1 boat.

but we can all boats put together with a "line" so when some guy sleep in 1 boat, the other guys move him forward.

The line is when apps can easy port by having same API.
The Linux World go this way and everybody can use the system he want.If that is maybe PPC or X86 or ARM or 68k or other.

On Linux i think 90% use on Desktop/Server X86, but nobody say other CPU are dead.All other CPU have the same API and the Linux devs spend many work to make porting easy by doing configure scripts.

And because the linux world make porting easy, thats the important for OS4 or MOS.
When unix world do same as MOS or OS4 devs, i think there was no MOS or OS4 possible, no Compiler, no OWB, and most the other apps for OS4 MOS that are Linux Ports

>Or maybe that in 4 years, you'll have finally understood that not everyone wants to >develop on a dead OS,

thats clear that not everybody want code for all OS, but that you call 68k AOS as a dead OS this propaganda i have hope change in next few years and the other sides notice the reality that all MOS OS4 68k AOS AROS are the state of a dead OS with near no market and programs.For 68k you can buy new cheap and fast X86 Hardware, and it is furtherdevelop, have actual compilers, have upto date Unix libs and good graphical debuggers to develop programs fast.

And thats i miss on the other systems, so i stay on 68k until the other Systems reach this minumum features.

And to make AOS more alive there help only working together.

Quote from: Karlos;537947
No need to knock C++ in this thread. It's perfectly possible to write fast C++ code. It's my preferred development language, even on m68k amigaos.



Yes its possible to write fast programs in C++ too, but you need then lots asm knowledge, and need look at the asm Output,  because you can easy write code a = b in 1 line that do thousands of CPU cycles.And because many of the C++ students do not know anything about how a CPU work, they write in most case not fast code.So it happen that C++ programs are big and slow and beginning and lots work is already flow in code optimizing to run fast enough, but fit not so good in the OO sheme.

I think c++ is good when we have 30 GHZ CPU 8 GB Ram SSD Harddisk that transfer 600 Megabyte /Sec, then speed is with clean OO Code really fast enough and also the compile speed of big c++ programs is not too large.

The main reason that i not try to port firefox now is because the long C++ compile time.
its really frustrating to do a change in a header file and must wait over 30 minutes until you can test it to see if you do all correct.And when you have not done correct you need wait another 30 minutes.

When i type make i dont want wait longer as 2-3 minutes for the working exe.

netsurf btw compile in 45 sec and so it make fun.
Title: Re: @Bernd_afa: OWB (68k) optimizations
Post by: wawrzon on January 13, 2010, 12:28:35 PM
i love this allegory with the sinking ship and boats.:D. by the way its not "line"; its "rope" ..

in fact i think bernd might seem boring like kato endlessly repeating his arguments, but he is right in a few things. one of them is, that i would really stop to call other solutions "dead systems" before one of these solutions actually prove itself to be succesful in a real world. till then they all can be perceived "dead".