Well, OWB is really great. But how is it doing compared to other common browsers?
I did a quick comparison with Opera on my netbook.

Reallife subjective tests:
I entered some urls and roughly measured the time until the page gets usable.
For n-tv.de, tagesschau.de, ebay.de, bahn.de and spon.de OWB indeed yielded to faster results than Opera. At tagesschau.de OWB delivered readable content after 2s while Opera took 4s. On ebay I was able to enter content to the search string after 2s while with Opera it took about 3s. bahn.de took 6s with OWB and a bit longer with Opera (difference not too big).
All n all reallife tests were in favor to OWB1.12/MorphOS.

Artificial tests:
SunSpider JS-benchark
- Opera: 1.1s
- OWB: 4.1s
peacekeeper benchmark
- Opera: total 1979 points (rendering: 1833, social networking: 1637, Complx graphics: 3073, Data: 1690, DOM: 1469, Text parsing: 4079)
- OWB: 687 points (rendring: 587, social networking: 712, Complx graphics: 1049, Data: 582, DOM: 1040, Text parsing: 605).

Well, the artificial tests looked bad for OWB in comparision with Opera. Surprising that the real life tests were resulting in quite other results. I guess a JS JIT would speed up things dramatically.

Test systems:
Mac mini G4/1500 OWB 1.12 MorphOS 2.7
Asus Eee 900A N270/1600 WinXP SP3 Opera 11.01

Systems hung on the same 100Mb router (no WLAN) connected to my 2Mb DSL by deutsche Telekom.

The N270/1600 and the 744x/1500 should be rather on par horsepowerwise.

All in all the netto browsing experince seems not to be slower with OWB than with Opera, albeit the theoretical numbers indicate a different expectation.

http://clients.futuremark.com/peacekeeper/index.action
http://www.webkit.org/perf/sunspider/sunspider.html

[ Editiert durch Zylesea an 2011/3/9 15:43 ]

I ran the artificial tests through TenFourFox 4.0b13pre on my 1.5 GHz G4 Mac mini with Leopard.

> SunSpider JS-benchark
> [...]
> - OWB: 4.1s

TenFourFox: 2.9s

> peacekeeper benchmark
> [...]
> - OWB: 687 points (rendring: 587, social networking: 712,
> Complx graphics: 1049, Data: 582, DOM: 1040, Text parsing: 605).

TenFourFox: 1049 points (rendering: 450, social networking: 768, complex graphics: 1445, data: 1670, DOM: 924, text parsing: 2392)

> I guess a JS JIT would speed up things dramatically.

My tests were done with JS JIT enabled.

Quote:

They don't look that bad if you take same HW (without JIT available for any of them), quite opposite. Here's my tests I made earlier:

1. OWB 1.12 / MorphOS 2.7 3909.3ms
2. Safari 4.1.3 / OSX 10.4.11 6171.0ms
3. Firefox 3.6.13 / OSX 10.4.11 8069.9ms
4. Opera 10.63 / OSX 10.4.11 8533.4ms

Full results.

Fortunately, most pages don't abuse javascript as extensively as these benchmarks, which is why it doesn't really show in practice.

But if you take the NES emulator coded in javascript, or other similar intensive programs, OWB won't have any chance against JIT-enabled browsers. :)

And there's also another area where we're behind: the hardware acceleration for the graphic rendering. The lack of it doesn't show that much during regular scrolling because only the strict necessary is being redrawn, but in applications like video players, or page with static backgrounds, you can easily see how slow software cairo can be.

Is hardware rendering being worked on ?

> peacekeeper benchmark
> [...]
> - OWB: 687 points (rendring: 587, social networking: 712,
> Complx graphics: 1049, Data: 582, DOM: 1040, Text parsing: 605).

OWB 1.13: 820 points (rendering: 630, social networking: 792, complex graphics: 1057, data: 1037, DOM: 1105, text parsing: 652)

19% overall improvement over OWB 1.12, nice :-)

> OWB 1.13: 820 points (rendering: 630, social networking: 792,
> complex graphics: 1057, data: 1037, DOM: 1105, text parsing: 652)

OWB 1.14: 873 points (rendering: 608, social networking: 764, complex graphics: 1050, data: 1025, DOM: 1090, text parsing: 979)

Again 6.5% faster overall :-)

That was quick!

to come out I mean.

Wonder if it fixes the eBay thinking OWB is a mobile browser or not? If that's anything to do with the program.

> Wonder if it fixes the eBay thinking OWB is a mobile browser or not?
> If that's anything to do with the program.

I get the normal site again but don't know if that's due to OWB or eBay.

So am I now, with ODYSSEY web browser 1.14, maybe eBay fixed itself.

Still can't see what was wrong with Origyn, grumble grumble..

1.14 Working well BTW!

[ Edited by stephen_robinson 06.08.2011 - 21:56 ]

The user-agent was changed a bit to prevent silly sites from thinking OWB is a mobile browser. :)

As for the name change, the point was to differentiate from the other OWB ports, especially when MorphOS OWB has almost nothing left from the original program.

> Still can't see what was wrong with Origyn

Some opinions:

https://morph.zone/modules/news/article.php?item_id=1877&comment_id=5344&mode=thread

Update:

> TenFourFox: 1049 points (rendering: 450, social networking: 768,
> complex graphics: 1445, data: 1670, DOM: 924, text parsing: 2392)

TenFourFox 7.0b1:
1093 points (rendering: 470, social networking: 798, complex graphics: 1313, data: 1613, DOM: 1110, text parsing: 2323)

Addendum:

>> OWB 1.13: 820 points (rendering: 630, social networking: 792,
>> complex graphics: 1057, data: 1037, DOM: 1105, text parsing: 652)

> OWB 1.14: 873 points (rendering: 608, social networking: 764,
> complex graphics: 1050, data: 1025, DOM: 1090, text parsing: 979)

Seems OWB 1.15 won't be even faster but in between 1.13 and 1.14:

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=34243&forum=32#628862

Well, be careful with this test, it can vary quite a lot depending on window size and so on. Besides, it's subject to change at each webkit revision, depending on javascriptcore state. The 1.15 version will surely use a more recent WebKit than the one that was used for this test.

> be careful with this test, it can vary quite a lot depending on window size and so on.

With OWB 1.12 I had the same score as reported by Zylesea here in this thread (on the same hardware), so I guess we must have used the same window size (mine is 1680x1050, minus some vertical lines for the screen bar) back then. If he didn't change window size meanwhile (I didn't) then the results should still be comparable.

> it's subject to change at each webkit revision, depending on javascriptcore state.
> The 1.15 version will surely use a more recent WebKit than the one that was used
> for this test.

Fine. So here I'm still hoping for OWB 1.15 to outdo OWB 1.14 in the Peacekeeper benchmark :-)

> http://clients.futuremark.com/peacekeeper/index.action

New Peacekeeper benchmark version available:

http://peacekeeper.futuremark.com

OWB 1.14:
265 points (HTML5 capabilities: 4/6, rendering: 9.8, HTML5 canvas: 1.6, data: 7432, DOM operations: 1590, text parsing: 7118)

TenFourFox 8.0b1:
374 points (HTML5 capabilities: 4/6, rendering: 6.5, HTML5 canvas: 2.7, data: 13306, DOM operations: 1777, text parsing: 17741)

The results on my Powermac G4 Gigabit (1.8GHz 7447a) with MorphOS 2.7 and OWB 1.14:

270 points (HTML5 Capabilities 4/6, Rendering 8.54, HTML5 Canvas 1.61, Data 7640.82, DOM operations 1661.05, Text parsing 8306.28)

[ Edited by analogkid 10.10.2011 - 08:26 ]

> OWB 1.14: 873 points (rendering: 608, social networking: 764,
> complex graphics: 1050, data: 1025, DOM: 1090, text parsing: 979)

OWB 1.15: 870 points (rendering: 612, social networking: 772, complex graphics: 1079, data: 1064, DOM: 1029, text parsing: 969)

> http://peacekeeper.futuremark.com
>
> OWB 1.14:
> 265 points (HTML5 capabilities: 4/6, rendering: 9.8, HTML5 canvas: 1.6,
> data: 7432, DOM operations: 1590, text parsing: 7118)

OWB 1.15:
241 points (HTML5 capabilities: 4/6, rendering: 9.8, HTML5 canvas: 1.6, data: 3316, DOM operations: 2067, text parsing: 7407)

> SunSpider JS-benchark
> [...]
> - OWB: 4.1s

TenFourFox 9.0b1: 2.1s

>> SunSpider JS-benchark
>> [...]
>> - OWB: 4.1s

> TenFourFox 9.0b1: 2.1s

TenFourFox 9.0 (MethodJIT enabled): 1.9s

>> http://peacekeeper.futuremark.com

> OWB 1.15:
> 241 points (HTML5 capabilities: 4/6, rendering: 9.8, HTML5 canvas: 1.6,
> data: 3316, DOM operations: 2067, text parsing: 7407)

OWB 1.16:
219 points (rendering: 8.6, HTML5 capabilities: 5/7, HTML5 canvas: 1.5, data: 3167, DOM operations: 1718, text parsing: 7370)

^^ Just checked, OWB 1.16 is out, on the same nite that YAM was updated, what are the odds! etc. etc.

http://fabportnawak.free.fr/owb/


It's like Christmas came early!!!

I'll get me coat.

[ Edited by stephen_robinson 26.12.2011 - 15:58 ]

> TenFourFox 7.0b1:
> 1093 points (rendering: 470, social networking: 798, complex
> graphics: 1313, data: 1613, DOM: 1110, text parsing: 2323)

TenFourFox 10.0b1:
1104 points (rendering: 487, social networking: 722, complex graphics: 1764, data: 2364, DOM: 1160, text parsing: 1704)