Google Chrome a Engineering Masterpiece ?

[CruNcher]

First thing that blow me away literaly was the start time on Windows (i couldn't belive it) it starts as fast as IE 7 if not subjectively even faster and it doesn't seem to be docked in the OS in anyway it easily beats Opera and Firefox in that point.

Tough right after the start it calls home no idea what it does yet (source code has to be checked) but it doesn't seem to be the Autoupdate (all Settings where on Private and no Search was active).


Some technical Points:
+ Really fast Startup time
+ Task Manager

Some Security Points:
+ Sandbox Concept (throughout the whole browsing except for Plugins)
+ Phising Check via Google (like Firefox)

Some usability UI points:
+ Tabs get smaller the more you add (Opera like)
+ Integrated Spell Checking
+ Most recent visited Sites visualy Displayed on ech new Tab (Opera like)
+ Maximum Screen area for Browsing (@ Default no Caption,Menus, only Adress and Tab Bar) all major functions accessable via the Adress Bar (2 popup icons)


Test: 36 Tabs + Video Playback 720p H.264

Memory Management (Video Playback not included):
http://s6.directupload.net/images/080904/rtu3js7q.png
http://s8.directupload.net/images/080904/mug532vz.png
http://s6.directupload.net/images/080904/9e2nuwef.png

Javascript execution (dromaeo.com)
Chrome = 619.40ms (Tota)
Firefox 3.1b1 = 2326.60ms (Total)
Opera = 3221.80ms (Total)

Acid3 (http://acid3.acidtests.org/)
Chrome = 79/100
Opera = 84/100
Firefox 3.1b1 = 85/100

Subjective Video Playback 720p Flash H.264 (http://somestuff.org/flashAVC/flvplayer.php?moviename=movies/BBB-1080downsized-x1280y720.mp4)
Opera = Some glitches
Chrome= Smoother
Firefox 3.1b1 = Unviewable extreme glitches

Bugs found so far (Chrome):
Doesn't paste @ the right place after copy when doing it with the right click Menu in a Text field (pastes @ end of text) ;)

Info: Chrome was the only browser running the whole time (memory leak possible) running all 3 @ once Firefox 3.1 creates random long lasting Kernel Explosions

[smok3]

i like it, but some cons

- seems like only one language is allowed thought (at the same time), this should be switched (as with firefox) at the current formarea.
- it's b**ugly
- font/html engine is not perfect.

i do see it as my default browser really soon thought.

p.s. this post written with;
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US) AppleWebKit/525.13 (KHTML, like Gecko) Chrome/0.2.149.27 Safari/525.13

[fibbingbear]

I have a couple of reasons why I don't want to use Chrome (yet):

1.) Lack of plugins.
2.) It's beta. Who knows what security holes are there (some have already been found since it uses an older build of WebKit). Other bugs have been found, like one that crashes every tab (even though Google claims it was designed to not do so).

Once these issues are addressed, it can be a serious contender. I think calling it an engineering masterpiece is a hyperbole, although it is well designed. It's also important to note that Firefox is planning an overhaul of its Javascript engine, so Chrome's speed (other than launch speed) may not be so big in the future.

One thing that I'm absolutely amazed by, though, is the DNS prefetching. I was surprised that other browsers didn't have this feature.

[LoRd_MuldeR]

I tried Chrome and it's really unstable at the moment ;)

* Click on "Options" causes "Woah! Google Chrome has crashed" (100% reproducible)
* Click on "Report bug or broken website" causes "Woah! Google Chrome has crashed" (100% reproducible)

Also the claims that Google Chrome is faster than Firefox/SeaMonkey are more than questionable:
Try "Streetview" in Google Maps. Runs pretty smooth in SeaMonkey 2.0, but runs pretty jerky in Chrome!

And why the hack does it install itself to this folder, without asking the user for the desired install location?
C:\Documents and Settings\John Doe\Local Settings\Application Data\Google\Chrome

Also any browser without Adblock Plus plugin is completely useless for my everyday work...

[smok3]

i have a problem with a process called 'google installer' which seems to call home often bothering my firewalls, however, design-wise the browser seems years ahead of firefox (didn't crash for me once yet, tested on two different machines).

p.s. about js speed, is there a sane way to benchmark this?

[LoRd_MuldeR]

i have a problem with a process called 'google installer' which seems to call home often bothering my firewalls

Delete/Rename "C:\Documents and Settings\LoRd_MuldeR\Local Settings\Application Data\Google\Update\GoogleUpdate.exe" and don't get bothered again :)

p.s. about js speed, is there a sane way to benchmark this?

Well, try one of those fancy Aajax applications. For example use Google Maps and test the "Street View" feature. Move the camera around.

Performance in SeaMonkey 2.0 is much better than in Chrome! Same should apply to Firefox 3.0/3.1 ...

[smok3]

ah, i see there is dromaeo.com

this is what i got;

http://dromaeo.com/
614.00ms.......chrome
3071.00ms......ff 3.0.1
3190.80ms......opera 9.52
----...........ie7 stoped responding after 1st test

so this is actually a significant improvement.

---

i have nuked the google updater, will see if thats it. :)

---

some things noticed:
- open a fat pdf in one tab, you can still click around the others (try that with ff)
- there is built-in drag-drop ability for the upload forms (you would need a plugin for that in ff)

[LoRd_MuldeR]

Here are my results:

Chrome 376.00 ms
SeaMonkey 2.0* 1182.00 ms
SeaMonkey 2.0 1574.60 ms
Opera 9.5 1468.40 ms
Safari 3.1.2 1665.80 ms
IE 8.0 2634.60 ms

(*) with JavaScript JIT enabled

Well, Chrome seems to be very well optimized for this kind of synthetic benchmarks. But what does it tell us about the performance on "normal" web-sites ??? :confused:
As sad said before, Chrome performance drops significantly on Google Maps StreetView (for example), while SeaMonkey is fine...

[smok3]

yes, i guess it does not tell much, however don't confuse ajax with javascript, slow ajax performance is much more than only js to blame.

[cyberbeing]

Firefox 3.1b1pre with JIT enabled beats out Chrome in the Sunspider JavaScript benchmark by 21% overall yet Chrome is somehow 270% faster in Dromaeo...

Firefox 3.1 JIT enabled (1725.8ms Total) (http://www2.webkit.org/perf/sunspider-0.9/sunspider-results.html?%7B%223d-cube%22:%5B57,57,55,57,57%5D,%223d-morph%22:%5B36,37,37,36,36%5D,%223d-raytrace%22:%5B79,77,77,90,79%5D,%22access-binary-trees%22:%5B59,60,60,82,57%5D,%22access-fannkuch%22:%5B160,156,156,159,159%5D,%22access-nbody%22:%5B41,42,43,41,42%5D,%22access-nsieve%22:%5B14,15,16,16,16%5D,%22bitops-3bit-bits-in-byte%22:%5B1,2,1,1,1%5D,%22bitops-bits-in-byte%22:%5B12,12,8,14,9%5D,%22bitops-bitwise-and%22:%5B11,11,11,11,12%5D,%22bitops-nsieve-bits%22:%5B29,29,28,29,28%5D,%22controlflow-recursive%22:%5B57,59,57,56,57%5D,%22crypto-aes%22:%5B39,39,37,38,38%5D,%22crypto-md5%22:%5B35,35,36,34,35%5D,%22crypto-sha1%22:%5B13,13,13,13,12%5D,%22date-format-tofte%22:%5B121,121,123,120,117%5D,%22date-format-xparb%22:%5B95,93,94,95,95%5D,%22math-cordic%22:%5B40,40,40,40,40%5D,%22math-partial-sums%22:%5B19,19,19,19,19%5D,%22math-spectral-norm%22:%5B8,8,8,9,8%5D,%22regexp-dna%22:%5B234,232,237,240,239%5D,%22string-base64%22:%5B18,18,18,17,18%5D,%22string-fasta%22:%5B91,92,91,95,92%5D,%22string-tagcloud%22:%5B117,118,118,114,122%5D,%22string-unpack-code%22:%5B274,260,268,261,281%5D,%22string-validate-input%22:%5B66,63,64,61,63%5D%7D)
Google Chrome (2192.0ms Total) (http://www2.webkit.org/perf/sunspider-0.9/sunspider-results.html?%7B%223d-cube%22:%5B32,30,31,31,34%5D,%223d-morph%22:%5B62,59,59,57,61%5D,%223d-raytrace%22:%5B42,43,43,43,43%5D,%22access-binary-trees%22:%5B7,7,7,5,8%5D,%22access-fannkuch%22:%5B38,37,37,37,37%5D,%22access-nbody%22:%5B31,33,34,33,32%5D,%22access-nsieve%22:%5B25,26,25,25,26%5D,%22bitops-3bit-bits-in-byte%22:%5B5,5,4,5,5%5D,%22bitops-bits-in-byte%22:%5B10,11,9,11,10%5D,%22bitops-bitwise-and%22:%5B25,27,29,30,27%5D,%22bitops-nsieve-bits%22:%5B32,32,33,32,31%5D,%22controlflow-recursive%22:%5B3,3,3,3,3%5D,%22crypto-aes%22:%5B26,25,24,24,26%5D,%22crypto-md5%22:%5B19,19,19,20,19%5D,%22crypto-sha1%22:%5B17,18,18,18,19%5D,%22date-format-tofte%22:%5B278,279,277,279,266%5D,%22date-format-xparb%22:%5B325,365,329,360,313%5D,%22math-cordic%22:%5B74,65,65,73,74%5D,%22math-partial-sums%22:%5B49,40,41,42,40%5D,%22math-spectral-norm%22:%5B18,16,18,17,17%5D,%22regexp-dna%22:%5B464,463,463,463,466%5D,%22string-base64%22:%5B73,72,73,77,75%5D,%22string-fasta%22:%5B67,69,68,63,61%5D,%22string-tagcloud%22:%5B165,174,163,175,162%5D,%22string-unpack-code%22:%5B215,213,217,218,222%5D,%22string-validate-input%22:%5B82,83,83,87,85%5D%7D)

Firefox JIT enabled (1247.80ms), Google Chrome (463.40ms) (http://dromaeo.com/?id=26298,26306)

[LoRd_MuldeR]

yes, i guess it does not tell much, however don't confuse ajax with javascript, slow ajax performance is much more than only js to blame.
That is right of course. But I care much more about the overall performance on Ajax sites than I care about the results for JavaScript-specific benchmarks...

These JavaScript-implemented games might be another test for JS performance:
http://www.nihilogic.dk/labs/javascript-games/

[lucassp]

* Click on "Options" causes "Woah! Google Chrome has crashed" (100% reproducible)
* Click on "Report bug or broken website" causes "Woah! Google Chrome has crashed" (100% reproducible)


Well...I reported a few broken web sites and it didn't crash.

I use Meebo all day long while I'm at work and it's very smooth on Chrome while on the other browsers isn't.

Of course, there's still lots of work to be done. But the Java Engine is great!

[smok3]

i have found some really obvious things borken, (try 32 bit png and set the CSS trasparency property to say 70%). Other than that png seems fine;
http://somestuff.org/png/png_test2.htm (my really old png test page)

http://somestuff.org/png/logo.png
(8bit rgba png - which is not something very usual)

[Sharktooth]

works here. even the png tests works.

[smok3]

those do work, try the transparency via css (on 32 bit png).
edit: quick and dirty example http://somestuff.org/png/png_test2_change.htm (varies transparency onmouseover)
it basically kills the original alpha, if transparency (other than default) is defined via css.

[Sharktooth]

ah, got it.

[Bigmango]

I tried Chrome and it's really unstable at the moment ;)

* Click on "Options" causes "Woah! Google Chrome has crashed" (100% reproducible)
* Click on "Report bug or broken website" causes "Woah! Google Chrome has crashed" (100% reproducible)

I don't have these problems. Chome has been rock stable here for several days now, on vista x64.


Also the claims that Google Chrome is faster than Firefox/SeaMonkey are more than questionable:
Try "Streetview" in Google Maps. Runs pretty smooth in SeaMonkey 2.0, but runs pretty jerky in Chrome!


I also don't have this problem, streetview is the same in firefox 3 and Chrome on my system (core2duo 2.13 ghz)

[Bigmango]

But what does it tell us about the performance on "normal" web-sites ??? :confused:


In my experience Chrome is MUCH faster with web applications. I.ex for me gmail, goodle docs, etc.... are real time with Chrome. Things are just there as soon as I click, it feels like I'm not connected to the internet, like these apps are installed on my PC. With firefox 3 it takes much longer to load things.

For normal websites it's about the same as firefox; some heavy forums are loading and refreshing faster with chrome.

[Blue_MiSfit]

I noticed a speed difference right away, even on doom9.

The javascript performance is nothing short of phenomenal :)

It feels surprisingly polished overall and quite usable.

Now, if only the ctrl+F popup could be relocated to the bottom (for firefox workalikeability :)), and if plugins could be ported (stumbleupon is all I care about ATM)

Also, they should lose the whole pokemon / simon hybrid icon concept. It's dumb.

Oh yeah, and they need to modify middle mouse click / drag scrolling behavior to mirror firefox / I.E.

I like feeling like the latency of my internet connection is more of a bottleneck than my PC / browser!

~MiSfit

[tetsuo55]

I have a couple of reasons why I don't want to use Chrome (yet):

1.) Lack of plugins.
2.) It's beta. Who knows what security holes are there (some have already been found since it uses an older build of WebKit). Other bugs have been found, like one that crashes every tab (even though Google claims it was designed to not do so).


1.) Plugins are provided through Gears, there hardly are any though
2.) All reported security holes are either already fixed, about to be fixed or non-working because of the sandboxing
the chromium source is being cleaned up for a merge with the WebKit bleeding edge

I tried Chrome and it's really unstable at the moment ;)

* Click on "Options" causes "Woah! Google Chrome has crashed" (100% reproducible)
* Click on "Report bug or broken website" causes "Woah! Google Chrome has crashed" (100% reproducible)

Also the claims that Google Chrome is faster than Firefox/SeaMonkey are more than questionable:
Try "Streetview" in Google Maps. Runs pretty smooth in SeaMonkey 2.0, but runs pretty jerky in Chrome!

And why the hack does it install itself to this folder, without asking the user for the desired install location?
C:\Documents and Settings\John Doe\Local Settings\Application Data\Google\Chrome

Also any browser without Adblock Plus plugin is completely useless for my everyday work...

-You are probably running a 64bit os, there is a know bug about that.

-Some users are having performance problems, you should report that with as much extra information as you can.

-Chromium can be installed anywhere

-Complain to the creator of adblock plus, Chrome(ium) supports plugins through the Gears interface

i have a problem with a process called 'google installer' which seems to call home often bothering my firewalls, however, design-wise the browser seems years ahead of firefox (didn't crash for me once yet, tested on two different machines).

p.s. about js speed, is there a sane way to benchmark this?


-Chromium does not have this

Firefox 3.1b1pre with JIT enabled beats out Chrome in the Sunspider JavaScript benchmark by 21% overall yet Chrome is somehow 270% faster in Dromaeo...

Firefox 3.1 JIT enabled (1725.8ms Total) (http://www2.webkit.org/perf/sunspider-0.9/sunspider-results.html?%7B%223d-cube%22:%5B57,57,55,57,57%5D,%223d-morph%22:%5B36,37,37,36,36%5D,%223d-raytrace%22:%5B79,77,77,90,79%5D,%22access-binary-trees%22:%5B59,60,60,82,57%5D,%22access-fannkuch%22:%5B160,156,156,159,159%5D,%22access-nbody%22:%5B41,42,43,41,42%5D,%22access-nsieve%22:%5B14,15,16,16,16%5D,%22bitops-3bit-bits-in-byte%22:%5B1,2,1,1,1%5D,%22bitops-bits-in-byte%22:%5B12,12,8,14,9%5D,%22bitops-bitwise-and%22:%5B11,11,11,11,12%5D,%22bitops-nsieve-bits%22:%5B29,29,28,29,28%5D,%22controlflow-recursive%22:%5B57,59,57,56,57%5D,%22crypto-aes%22:%5B39,39,37,38,38%5D,%22crypto-md5%22:%5B35,35,36,34,35%5D,%22crypto-sha1%22:%5B13,13,13,13,12%5D,%22date-format-tofte%22:%5B121,121,123,120,117%5D,%22date-format-xparb%22:%5B95,93,94,95,95%5D,%22math-cordic%22:%5B40,40,40,40,40%5D,%22math-partial-sums%22:%5B19,19,19,19,19%5D,%22math-spectral-norm%22:%5B8,8,8,9,8%5D,%22regexp-dna%22:%5B234,232,237,240,239%5D,%22string-base64%22:%5B18,18,18,17,18%5D,%22string-fasta%22:%5B91,92,91,95,92%5D,%22string-tagcloud%22:%5B117,118,118,114,122%5D,%22string-unpack-code%22:%5B274,260,268,261,281%5D,%22string-validate-input%22:%5B66,63,64,61,63%5D%7D)
Google Chrome (2192.0ms Total) (http://www2.webkit.org/perf/sunspider-0.9/sunspider-results.html?%7B%223d-cube%22:%5B32,30,31,31,34%5D,%223d-morph%22:%5B62,59,59,57,61%5D,%223d-raytrace%22:%5B42,43,43,43,43%5D,%22access-binary-trees%22:%5B7,7,7,5,8%5D,%22access-fannkuch%22:%5B38,37,37,37,37%5D,%22access-nbody%22:%5B31,33,34,33,32%5D,%22access-nsieve%22:%5B25,26,25,25,26%5D,%22bitops-3bit-bits-in-byte%22:%5B5,5,4,5,5%5D,%22bitops-bits-in-byte%22:%5B10,11,9,11,10%5D,%22bitops-bitwise-and%22:%5B25,27,29,30,27%5D,%22bitops-nsieve-bits%22:%5B32,32,33,32,31%5D,%22controlflow-recursive%22:%5B3,3,3,3,3%5D,%22crypto-aes%22:%5B26,25,24,24,26%5D,%22crypto-md5%22:%5B19,19,19,20,19%5D,%22crypto-sha1%22:%5B17,18,18,18,19%5D,%22date-format-tofte%22:%5B278,279,277,279,266%5D,%22date-format-xparb%22:%5B325,365,329,360,313%5D,%22math-cordic%22:%5B74,65,65,73,74%5D,%22math-partial-sums%22:%5B49,40,41,42,40%5D,%22math-spectral-norm%22:%5B18,16,18,17,17%5D,%22regexp-dna%22:%5B464,463,463,463,466%5D,%22string-base64%22:%5B73,72,73,77,75%5D,%22string-fasta%22:%5B67,69,68,63,61%5D,%22string-tagcloud%22:%5B165,174,163,175,162%5D,%22string-unpack-code%22:%5B215,213,217,218,222%5D,%22string-validate-input%22:%5B82,83,83,87,85%5D%7D)

Firefox JIT enabled (1247.80ms), Google Chrome (463.40ms) (http://dromaeo.com/?id=26298,26306)

-Chromium SVN should beat those scores

i have found some really obvious things borken, (try 32 bit png and set the CSS trasparency property to say 70%). Other than that png seems fine;
http://somestuff.org/png/png_test2.htm (my really old png test page)

http://somestuff.org/png/logo.png
(8bit rgba png - which is not something very usual)

This is a known bug, maybe you could help fix it?

[LoRd_MuldeR]

-You are probably running a 64bit os, there is a know bug about that.

Yes, WinXP x64. But x64 Windows is fully compatible to x86 Windows. 32-Bit versions Firefox, SeaMonkey, Opera and Safari have never complained ;)

-Some users are having performance problems, you should report that with as much extra information as you can.

How? It will crash every time when I try "Report bug or broken website" and I cannot find a similar link on their web-site... :D

-Chromium can be installed anywhere

How? Once you double-click the installer.exe, it will start to download/install. Never asks the user for anything, except for accepting the license... :rolleyes:

-Complain to the creator of adblock plus, Chrome(ium) supports plugins through the Gears interface

Well, we will see how good the "community support" for Chrome will be. At the moment Firefox/SeaMonkey has got the most active community...

[tetsuo55]

Its a bug that only occurs when the user has a 64bit os, they will fix it eventually.

-------

You can report bugs manually here:
http://code.google.com/p/chromium/issues/list

-------

You can download chromium as a zip file here:
http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/

Unzip wherever you want

-------

Im guessing comunity support will eventually be huge:

-Gears plugins work on ALL browsers
-All the internet pioneers are working on Chrome(that includes the guys who made firefox)
-Chrome beats every other browser on the market even in its 0.2 beta state.

I personally don't care about ad's and i don't feel the browser is responsible for blocking them.

[CWR03]

One thing I noticed right away is that the mouse wheel scroll is inverted and ignored my "full page at a time" settings.

[Sharktooth]

latest chromium builds can be found here: http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/
remember they're testing builds. remove google chrome before installing those ones.

[LoRd_MuldeR]

latest chromium builds can be found here: http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/
remember they're testing builds. remove google chrome before installing those ones.

The latest Chromium build fixed the Options crash for me :)

[tetsuo55]

The latest Chromium build fixed the Options crash for me :)

Cool accidentally fixed :D

[Sharktooth]

firefox 3.1 alpha 2 is ready too.
windows (http://download.mozilla.org/?product=shiretoko-alpha2&os=win&lang=en-US), mac (http://download.mozilla.org/?product=shiretoko-alpha2&os=osx&lang=en-US) and linux (http://download.mozilla.org/?product=shiretoko-alpha2&os=linux&lang=en-US)

[smok3]

if it is a known bug, then what exactly should i do to get it fixed? (the png thingy i mean)

tnx for the chromium builds.

[Sharktooth]

wait for a fix or code it yourself possibly sending the patch to the chromium devs

[smok3]

and i'am a browser developer since when? :)

[Sharktooth]

you asked "what exactly should i do..." and i answered what you should do. maybe you missed the OR between "wait for a fix" and "code it yourself..."

[smok3]

yes, there is a difference between or and and, i'll wait, the bug was reported via the chrome.

[Astrophizz]

As for blocking ads in Chromium, this supposedly works: http://www.privoxy.org/

[buzzqw]

i prefer proxomitron , with sidki rules ( http://www.geocities.com/sidki3003/prox.html ), and it's working too with chrome

BHH

[Sharktooth]

in ver 2105 the png bug is still NOT fixed.
i guess this is a problem in webkit (can anyone with konqueror test it?) so the fix may require a while.

[LoRd_MuldeR]

This free tool is able to kill your personal Chrome ID and thus will prevent Chrome from sending user-specific info to Google servers:
http://www.almisoft.de/?cont=kchrome

[Sharktooth]

german...

[LoRd_MuldeR]

german...

Click the "Kill-ID" button on the left side and then "Alles deaktivieren" (disable everything) on the bottom and you are done :)

Things that will be disabled:
* Browser-ID
* send autocomplete suggestions to Google
* send none-existing (invalid) URL's to Google
* send user-statistics and crash logs to Google
* Google Updater

[smok3]

in ver 2105 the png bug is still NOT fixed.
i guess this is a problem in webkit (can anyone with konqueror test it?) so the fix may require a while.

there is no bug in safari.

[Sharktooth]

uhm... im downloading webkit

EDIT: ok, it's not a webkit issue

[tetsuo55]

Chrome uses a very old version of webkit (safari 3.0 ish)

There are a bunch of bugreports about the image problems. I saw one commit that fixes some of them. but obviously not the problem you are having.

They are trying to fix the bug and are going to upgrade webkit to TOT (which is even newer than the Safari 4 beta)

[Sharktooth]

build 2264. still the png problem, so obviously webkit was not updated...

[tetsuo55]

build 2264. still the png problem, so obviously webkit was not updated...

I guestimate the merge might take a lot of time.
They have reached a lot of milestones but they are not quite there yet (they changed to much to simply copy and paste over webkit)

[mr soft]

Just reading this, thought it was worth sharing .

Chrome is a security nightmare, indexes your bank accounts


http://www.tgdaily.com/content/view/39176/108/

[tetsuo55]

That bug/feature has been fixed in chromium for some time now

[smok3]

arh well, just went to sp3 and it wont start anymore :/

[Sharktooth]

reinstall it... works here with SP3.

[Sharktooth]

that damn thing is getting faster and faster...
there's still the png problem but rev. 2584 is outperforming every other browser i tried.
anyone knows if there is a developers blog or journal?

[LoRd_MuldeR]

I noticed a rendering bug in Chrome, which might be related to the PNG bug:

http://img510.imageshack.us/img510/601/chromebugpe4.th.png (http://img510.imageshack.us/my.php?image=chromebugpe4.png)http://img510.imageshack.us/images/thpix.gif (http://g.imageshack.us/thpix.php)

[Sharktooth]

definatly a rendering bug. the image didnt get resized before being displayed.
you can send bugreports at dirhael AT gmail DOT com.

[LoRd_MuldeR]

you can send bugreports at dirhael AT gmail DOT com.

Done.

[Dirhael]

definatly a rendering bug. the image didnt get resized before being displayed.
you can send bugreports at dirhael AT gmail DOT com.
Oops, I think there has been a misunderstanding here as I don't have anything to do with Chrome or Chromium...all I've made is that "nightly" updater utility. Guess it would be best if I changed/clarified the statusbar text for the next release ;)

[smok3]

reinstall it... works here with SP3.

chrome or chromium?

[LoRd_MuldeR]

Iron is the first(???) Chrome/Chromium clone and it comes without the Client-ID feature:
http://www.srware.net/software_srware_iron.php

It seems to be based on Chromium Build 1880 and it looks identical to the official Chrome...

[Sharktooth]

Oops, I think there has been a misunderstanding here as I don't have anything to do with Chrome or Chromium...all I've made is that "nightly" updater utility. Guess it would be best if I changed/clarified the statusbar text for the next release ;)
oops, sorry i tought that email address was for collecting chromium bugs...

@lord_mulder: considering the last build is 2617 (right now) it seems quite outdated

[LoRd_MuldeR]

Latest builds got Mouse Button-3 (press Wheel down) support for scrolling, finally :)

[Sharktooth]

it would be interesting to test the latest chromium builds vs the latest webkit builds with squirrelfish extreme since it seems squirrelfish xt is even faster than V8

[CruNcher]

WOW all Win32 users look @ the new Firefox 3.1b2 nightlies the Speed improved alot lately :D it's up to when not even faster in rendering then Google or Opera (subjectively) :) Doom9 renders like hell fast, everything fells a lot smoother now (going through the 36 tab test is now fun) seems they found the main things that where slowing it down under Windows the whole time (also Cache/Memory behavior seems to have been improved less i/o is going on hard drive doesn't sound wise explodes anymore) i find that a major subjective speed improvement (since a long long time) :)
The magic started @ 20081007 acidtest3 is now @ 88% Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b2pre) Gecko/20081009 Minefield/3.1b2pre (.NET CLR 3.5.30729) ID:20081009033154
Though i was keeping more an eye on Firefox lately have to check how Google Chromium improved though Opera didn't made any big steps (it was already one of the most performant under Win32 seems it's slowly getting now @ the end and Google and Firefox drive away :D)

[smok3]

well, i'am certainly back to firefox, speed is not everything.

[Sharktooth]

FF is still too slow to start up. chrome/chromium is blazing fast.
Opera is fast to, not as chrome/chromium though.
Also Opera has not a maniacally optimized JS engine (and it already is one of the fastest browsers on earth), so there is still a lot of room for improvements.

[CruNcher]

Can't agree since the last nightlies also startup is really fast though i also deactivated all the google site security stuff (malware/trusted sites) scanning (sql database) maybe that has todo with it hmm
http://dromaeo.com/?id=46705 <- newer nightlies currently stop @ the dom test :(
acidetest3 = 89

[LoRd_MuldeR]

Here is my attempt to create an Auto Updater for Chromium:
http://www.mediafire.com/file/dzmywjkmzdz/Chromium-Updater.2008-10-17.Try-4.zip

It will check for the latest build, download the ZIP file and extract it for you...

[Astrophizz]

Looks like they're keeping ahead of Firefox 3.1 in speed: http://news.cnet.com/8301-17939_109-10068743-2.html?part=rss&tag=feed&subj=Webware

[LoRd_MuldeR]

Here is my attempt to create an Auto Updater for Chromium:
http://www.mediafire.com/file/dzmywjkmzdz/Chromium-Updater.2008-10-17.Try-4.zip

It will check for the latest build, download the ZIP file and extract it for you...

Here is a minor update:
http://www.mediafire.com/file/15zncnzdj34/Chromium-Updater.2008-10-19.zip

It will check for running instances of Chrome now!

[Bigmango]

Looks like they're keeping ahead of Firefox 3.1 in speed: http://news.cnet.com/8301-17939_109-10068743-2.html?part=rss&tag=feed&subj=Webware

Actually, in the link you have posted, Firefox is the fastest. Tracemonkey is the new java engine coming with Firefox 3.1.

"...bear in mind here that a smaller score is best for SunSpider: TraceMonkey-enabled Firefox led with a score of 2,257; Chrome was second at 2,904..."

The google javascript test is currently not relevant until the Firefox bug is fixed.

[Astrophizz]

Good point, thanks for correcting me.

[Bigmango]

Good point, thanks for correcting me.

Anyway, the difference is very small. They are both as fast. :)

[Astrophizz]

Yeah, the progress with this stuff right now makes me think that there really wasn't much focus on speed in years past - except in Opera. Or maybe the new javascript engines are some sort of breakthrough.

[LoRd_MuldeR]

Another small update to my Chormium Auto-Update tool:
http://www.mediafire.com/file/oovczfm5ttv/Chromium-Updater.2008-10-21.zip

The files will now be copied to the root folder instead of the "chrome-win32" subfolder.
So you can have "Chromium-Updater.exe" and "Chrome.exe" in the same folder now.

[LoRd_MuldeR]

Another small update to my Chormium Auto-Update tool:
http://www.mediafire.com/file/oovczfm5ttv/Chromium-Updater.2008-10-21.zip

The files will now be copied to the root folder instead of the "chrome-win32" subfolder.
So you can have "Chromium-Updater.exe" and "Chrome.exe" in the same folder now.

Yet another update to reduce the Spam on the installer's console:
http://www.mediafire.com/file/o3zmtndzjdy/Chromium-Updater.2008-10-24.zip

[LoRd_MuldeR]

Does anybody else make the experience that current Chromium builds crash on startup?

For me it started yesterday and it still happens with the build I just downloaded one minute ago...


[EDIT]

Update: It seems build 4855 is good again ;)

[smok3]

i always get this
http://blog.somestuff.org/images/iron_error.png
any clues?
(this happens with chrome, chromium, iron, system is xp pro sp3)

p.s. i only wanna use this browser for testing flash video playback really (at this point in time).

[LoRd_MuldeR]

i always get this
http://blog.somestuff.org/images/iron_error.png
any clues?
(this happens with chrome, chromium, iron, system is xp pro sp3)

p.s. i only wanna use this browser for testing flash video playback really (at this point in time).

Are you sure "compatibility mode" isn't used for chrome.exe ?

[smok3]

98% sure.

[LoRd_MuldeR]

98% sure.

Better make it 100% sure. I once enabled "compatibility mode" and forgot about it. It will make Chrome/Chromium crash at startup!

Also make sure you get the latest Chromium builds. A short while ago, all builds were broken for ~24 hours...

[smok3]

rmb click on icon and uncheck that sort of mode right? any other hidden variations? (in xp i mean)

[LoRd_MuldeR]

rmb click on icon and uncheck that sort of mode right?

That's what I'd do, yes.

[smok3]

found it, it appears the problem is connected with symantec, the workaround is to run the app like
iron.exe --no-sandbox

[LoRd_MuldeR]

found it, it appears the problem is connected with symantec, the workaround is to run the app like
iron.exe --no-sandbox

Good to know! You should report your finding to the Chromium guys, if you didn't do so yet...

[Sharktooth]

did anyone test if there were improvements in speed in the latest builds?
i have the impression it became faster for regular browsing but i dont know "when" it started...

[Shinigami-Sama]

found it, it appears the problem is connected with symantec, the workaround is to run the app like
iron.exe --no-sandbox

how does this not surprise me that they can't get it right?

[smok3]

it is under known issues and filled in as fixed
http://www.google.com/support/chrome/bin/static.py?page=known_issues.cs
but obviously is not fixed at all, will have to wait :)

[Sharktooth]

recent chromium builds score 100/100 on acid 3 test, but it's still not completely passed.
JS execution speed is improved as well but still there are some issues with DOM.
things seem to be moving at an impressive rate...

[Astrophizz]

Wow, that is pretty fast. I remember Opera devs got a 100/100 but didn't pass the time part (smooth animation). Haven't heard anything about it since and probably won't until they start to release more pre-10.0 builds.

[LoRd_MuldeR]

recent chromium builds score 100/100 on acid 3 test, but it's still not completely passed.

I just ran Acid3 with latest Chromium build (r6035) and it hangs at 66/100 with "LINKTEST FAILED", then moves on to 96/100 and stops there.

[Sharktooth]

when i wrote it, it was build 58xx (where xx = something i cant remember)

[LoRd_MuldeR]

when i wrote it, it was build 58xx (where xx = something i cant remember)

So they broke it again? :D


Here is my result from Chromium r6107:

http://img167.imageshack.us/img167/4285/chromeacid3hv1.png

[Taurus]

hmmmh, its working:
http://img175.imageshack.us/img175/9655/capdp7.th.png (http://img175.imageshack.us/my.php?image=capdp7.png)

[Sharktooth]

it failed just the linktest. same was when i tested it.
probably they're still working on it.

[fakey]

scores 97 with bfilter enabled, 100 with bfilter bypassed.
seems like the soft-proxy partially breaks something out of acid3.

[Sharktooth]

a big competition is on going...
opera software is about to release the first alpha of opera 10 with a faster presto 2.2 engine... mozilla is readying firefox 3.1 with trace monkey engine... next webkit (so safari) will get a massive speedup using the new squirrelfish (xtreme?) engine... google has chrome/chromium with V8 engine that's performing better and better... microsoft is fighting back with IE8...
some comparisons are here: http://arstechnica.com/journals/apple.ars/2008/09/19/squirrelfish-extreme-mops-floor-with-v8-tracemonkey
gentlemen, fasten the seatbelts, the race begins...

[CruNcher]

i like this really great stuff you really realize the speed boost especially on older systems it's great :) seeing that all are optimizing low level performance currently, but many people are only talking about javascript you should realize the speed increase in the whole rendering process that takes place (more efficient I/O and memory management). Yes Presto 2.2 sounds promising the Opera Labs last official release was a new test version of the Video embedded 9.62 http://labs.opera.com

[Sharktooth]

yup, html5 is going to change the web... :)
btw, that 9.62 build with file i/o and html 5 video tag is not based on presto 2.2.

[CruNcher]

http://www.opera.com/browser/next/


Edit: Imho Opera isn't that impressive anymore speed and optimized wise like it was some years ago (in the time of Firefox 1-2) Firefox (3.1 beta 3) pretty much has the same rendering Performance by now on Windows :)

Though im concentrating now more on Multimedia (Streaming,Playback) Performance and that is interesting how different they are here :)

[Sharktooth]

opera software headed the new generation of browsers... they always did.
they were always the first to develop and integrate new features we can now find in major web browsers.

[Sharktooth]

chrome is out of "beta": http://www.tcmagazine.com/comments.php?shownews=23403&catid=3

[LoRd_MuldeR]

Opera 10.00-alpha does not complete score 100/100 in the Acid3 test, although the homepage claim it does:

http://img247.imageshack.us/img247/8027/opera10ru8.th.png (http://img247.imageshack.us/img247/8027/opera10ru8.png)

[tetsuo55]

Opera 10.00-alpha does not complete score 100/100 in the Acid3 test, although the homepage claim it does:

http://img247.imageshack.us/img247/8027/opera10ru8.th.png (http://img247.imageshack.us/img247/8027/opera10ru8.png)

although opera does fix compatibility with a lot of websites, it still doesn't show many correctly

Chromium shows some sites that other browsers dont

[CruNcher]

Firefox got another Rendering Speedup recently, seems bug fixes where the reason for that :)
3.1 Final will be great especialy Windows user will realize a major Speed difference :)

[Sharktooth]

google chrome 3.0 is out (for devs): http://dl.google.com/chrome/install/182.2/chrome_installer.exe
it comes with a new WebKit (531.0), V8 (1.2.5.1) and Gears (0.5.21.0) versions, a number of crash fixes, support for video tags, a 'cleaner' UI, and more.
chromium is, as usual, a step ahead (183.0).
it would be interesting to compare the performance with safari 4 (beta), opera 10 (beta), firefox 3.1 (beta 4) and IE 8

[Sharktooth]

chromium for linux and osx is coming. google released test builds for developers...

[laserfan]

Hey does anyone here use Chrome day-to-day? I find it to be wicked-fast compared to my usual Avant-over-IE browsing, but a key barrier to full-time use is in browsing forums like doom9, i.e. with my usual "do=getnew" entry bookmarks to forum sites like this one, Chrome just dumps me to the bottom of the page. I have to scroll-up to see the newest posts. With IE a getnew puts you right at the newest post from the last time you visited.

Anyone here have the same irritation, or know of how to get it to work "correctly"?

[tetsuo55]

I don't experience that. are you using the latest build from the buildbot?

[laserfan]

I don't experience that. are you using the latest build from the buildbot?Hmmm, yeah 3.0.182.2 is the very latest I think. All builds to present have done this for me (getnew just dumps me to very bottom of thread page). Seen it on a couple of computers, both w/XP Pro 32-bit...

I haven't fiddled with any Options but don't see anything relevant in there anyway.

Thanks tetsuo55; does it work for everyone else here too???

[tetsuo55]

No wonder, you have an ancient build.

Get the latest build here (highest number at the bottom)
http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/

Yours is months old, and i remember that bug now, has been fixed.

[laserfan]

No wonder, you have an ancient build.Dunno how I could get left behind, I just followed instructions way-back-when to install, then it appeared to be updating itself periodically. When the heck did Chrome become Chromium anyway? No matter...

I downloaded the newest build and the Chromium therein (v3.0.197.0 build 22158) and it works as it should! Thanks tetsuo55 for your help!
:thanks:

[Astrophizz]

Chrome didn't become Chromium. He linked you to Chromium but you could stay with Chrome if you want. You should probably redownload chrome and set it to the developer channel if so (I had the problem once where Chrome stopped updating and I just had to reinstall it). Chromium won't autoupdate.

[LoRd_MuldeR]

When the heck did Chrome become Chromium anyway?

The Chromium browser is developed as an OpenSource project under a BSD license. The official releases (possibly containing proprietary extensions) are re-branded to "Chrome".

Chromium won't autoupdate.

Really? :D

http://mulder.dummwiedeutsch.de/home/?page=projects#chromium

[Astrophizz]

I use that, but that's not an autoupdate, at least not like what there is with Chrome. You have to manually run it.

[LoRd_MuldeR]

I use that, but that's not an autoupdate, at least not like what there is with Chrome.

Right. According to the Chrome license terms, the user must accept any Update that Google is sending (silently). That's a difference indeed ;)

You have to manually run it.

You could link it to Autostart or use the Windows Scheduler...

[sjakke]

Thanks for that updater LoRd_MuldeR. Works like a charm:)

By the way I was wondering if Chromium suffers from the same privacy problems as http://www.srware.net/en/software_srware_iron_chrome_vs_iron.php claims Chrome has?

[LoRd_MuldeR]

By the way I was wondering if Chromium suffers from the same privacy problems as http://www.srware.net/en/software_srware_iron_chrome_vs_iron.php claims Chrome has?

I think so. Otherwise "Iron" would be useless.

But since Chromium is OpenSource (in contrast to "official" Chrome releases), we can do whatever modifications we like (e.g. remove the User-ID feature).

And that's exactly what Iron does. However they don't update as often as the Chromium build bot does ;)

[sjakke]

Well I would like to continue, using Iron, but it's started crashing every time I open a page.

And, as you mentioned, they don't seem to do updates very often. So I've switched to Chromium.

[laserfan]

The Chromium browser is developed as an OpenSource project under a BSD license. The official releases (possibly containing proprietary extensions) are re-branded to "Chrome".Thanks, I had assumed as much--bizarre how hard it's been to find this out (between Chrome sites and Chromium sites that is). Good 'ol doom9.org to the rescue.

I'll stick with Chromium myself--what with Chrome's wierd "install per user" installation, and no release notes anywhere as to what it is your downloading, and having to download for every user, bah.

[LoRd_MuldeR]

I'll stick with Chromium myself--what with Chrome's wierd "install per user" installation, and no release notes anywhere as to what it is your downloading, and having to download for every user, bah.

Yeah, I hate those installers that install to %APPDATA% without asking for an install folder.

Obviously their intention is to allow non-admin users to install the application without an "evil" UAC dialog popping up on Vista/Win7 ;)

But this undermines the concept of having different folders for different purposes with different user access rights....

[Shinigami-Sama]

to me its still useless untill I can force it to load all tabs in the background...
switching to a new tab every time its spawned is a PITA, and no ctrl+shift+click isn't ganna cut it...

[laserfan]

to me its still useless untill I can force it to load all tabs in the background...
switching to a new tab every time its spawned is a PITA, and no ctrl+shift+click isn't ganna cut it...Mine doesn't work that way; at least, I usually right-click on a link and the tab is generated in the background...

[Shinigami-Sama]

Mine doesn't work that way; at least, I usually right-click on a link and the tab is generated in the background...

I use middle click and it automatically throws me into the new tab, I've found that if I ctrl+shift+click it goes in the background

I'm also one of those tab whores who has about 12 tabs open at the minimum

[tetsuo55]

Shinigami-Sama you have to update to the latest buildbot version.

[Midzuki]

Get the latest build here (highest number at the bottom)
http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/



I've tried to download from the URL above, but the server always interrupts the file transfer around 15%. Is there an explanation for the anomaly (other than "disliking IP addresses from the 3rd World" :) )?

[Shinigami-Sama]

Shinigami-Sama you have to update to the latest buildbot version.

not interested in testing again now that I solved the problem that made me try chrome in the first place

[laserfan]

I use middle click and it automatically throws me into the new tab...Isn't that funny--I've been using "personal computers" since before they were called that, and before the "mouse" was even introduced (mid-70s) and never tried "middle-clicking"!

Works fabulous, no more right-click and select "open in new tab"! Opens in background too! Thanks Shinigami-Sama! :)

[Midzuki]

OK, finally managed to download a two-week-old version from a Tucows mirror --- that sucks, really. :(

My first (and last) impression: better than the Chrome I tested some months ago, but still uglier than IE. :p Back to the olde and goode K-Meleon. :D

[laserfan]

Apart from the printing oddities (can't change header/footer or margins) it is just wildly, ridiculously faster than IE, at least with my ISP. I'm kinda in awe of it as it's made my satellite service surfing seemingly a ROM faster, instantly. I've been using Avant on top of IE.

[LoRd_MuldeR]

Apart from the printing oddities (can't change header/footer or margins) it is just wildly, ridiculously faster than IE

That applies to any browser, except IE and IE-based ones :D

IE 8.0 would be the very first IE to be competitive after years if it wasn't that unbearably slow :rolleyes:

[Midzuki]

On my obsolete Pentium 4, the fastest and most-elegant web browser still is K-Meleon. :) The vice-champion is Opera, and the third place goes to Seamonkey. :p Firefox still smells like Netscape 6 to my nose, and I cannot drop IE8 altogether, because there are web sites which only can be viewed correctly with the Trident engine. :mad:

[LoRd_MuldeR]

The vice-champion is Opera, and the third place goes to Seamonkey. :p

SeaMonkey 2.0 Nighly is my primary browser :)

I cannot drop IE8 altogether, because there are web sites which only can be viewed correctly with the Trident engine. :mad:

I'm yet to encounter such a web-site. Anyway, I'd simply add that site to my ignore list :p

[tetsuo55]

I'm yet to encounter such a web-site. Anyway, I'd simply add that site to my ignore list :pI know a ton of these unfortunately, and ignoring them is not an option.

sites like my bank, government, etc...

[LoRd_MuldeR]

I know a ton of these unfortunately, and ignoring them is not an option.

sites like my bank, government, etc...

Too bad. Still unless people start to complain about that problem, they won't even notice or not feel responsible to fix it...

[tetsuo55]

they refuse to fix it when 98+% of users use a flavor of IE

[LoRd_MuldeR]

they refuse to fix it when 98+% of users use a flavor of IE

Not exactly :D

http://img199.imageshack.us/img199/7355/browserstats.png

Maybe my little web-site isn't representative for the entire web. But the results shouldn't be completely misleading either ;)

Furthermore, standards like HTML, JavaScript and CSS exist for a reason. If web designers followed these standards, it would work in any browser. If not, it's their mistake!

[tetsuo55]

Sorry, i meant the visitors of those site (because i reported the incompatibilities years ago, the sites dont work on any browser other than IE)

[LoRd_MuldeR]

Sorry, i meant the visitors of those site

If your web-site simply doesn't work with anything else but IE, it's not hard to have 98% IE users in your site statistics :p

i reported the incompatibilities years ago, the sites dont work on any browser other than IE

Good. If more users did, they probably would see that it's necessary to fix the problem. Unfortunately I think most non-IE users sigh, load up IE and proceed...

[tetsuo55]

thats what i do now (no sense in reporting every time) :(

[smok3]

maybe another approach would work: 'just make it work for firefox, each new ie is getting slightly better, so old pages are looking better and better (for those XY% of IE users)'.

edit: the pie for my blog
http://blog.somestuff.org/images/browsers.png
pretty sad...

[Midzuki]

As tetsuo55 said, many bank and gov't sites are designed by IE fans :( . I could also mention certain p0rn sites, but I don't want to violate the forum rules. :) Anyway, and back to the topic:

Chrome and Chromium are not aimed at the so-called "power users". In Netscape 4.xx, I could disable the automatic loading of images for all the pages, and when I wanted to view the images of a particular page, I simply had to click the appropriate button on the toolbar. :) Or just right-click on the selected picture(s). In K-Meleon, I press F7, and the page is re-interpreted without the interference of the JavaScript code, and without having to reload the page. :devil: I don't know how to do the same "tricks" in Chromium (assuming that it does have such features, and they are just «very-well hidden» from the end-user).

[IgorC]

From what I see even Firefox is importing chromium code in this moment.
http://hg.mozilla.org/projects/electrolysis/rev/6fd4bb500d42

Firefox release version of 3.6 should have performance on par with Chrome 3.0. And 3.7 will be chromized even visually. :)

[Midzuki]

From what I see even Firefox is importing chromium code in this moment.
http://hg.mozilla.org/projects/electrolysis/rev/6fd4bb500d42

Only the good parts, let's hope. :p

[RNiK]

Only the good parts, let's hope. :pMozilla is using Google Chrome code to accomplish this (https://wiki.mozilla.org/Content_Processes). ;)

[smok3]

very funny?, especially since chrome will run on nixes very soon as it seems... ;)

Mozilla/5.0 (X11; U; Linux i686; en-US) AppleWebKit/532.0 (KHTML, like Gecko) Chrome/3.0.197.11 Safari/532.0

writing this on linux mint with chrome.

[Keiyakusha]

Does anyone knows how fast built-in h264 decoding in Chrome compared to Flash?

[RNiK]

Does anyone knows how fast built-in h264 decoding in Chrome compared to Flash?No, but indeed it would be an interesting comparison.

[smok3]

Does anyone knows how fast built-in h264 decoding in Chrome compared to Flash?

how would one benchmark that?