Surfin' Safari

Full Pass of Acid3

Posted by Maciej Stachowiak on Thursday, September 25th, 2008 at 5:29 pm

Today we would like to announce that WebKit is the first browser engine to fully pass Acid3. A while back, we posted that we scored 100/100 and matched the reference rendering. Now, thanks to recent speedups in JavaScript, DOM and rendering, we have passed the third condition, smooth animation on reference hardware.

Here is a screenshot of a successful run:

Here is the timing reference dialog you get by clicking on the “A” in Acid3 that confirms we pass the smooth animation condition on a 2.4GHz MacBook Pro:

To try it for yourself, grab a nightly. Keep in mind that on slower machines, the timing may not be perfect, and you need to do a cached run of the test (load it once, close window, open new window, load it again) to avoid delays from the network.

32 Responses to “Full Pass of Acid3”

  1. Oliver Says:

    Woo!

  2. jmah Says:

    According to http://ln.hixie.ch/?start=1207096078&count=1 , the test should complete in 33ms, but your dialog shows 330ms…

  3. Maciej Stachowiak Says:

    The individual tests (of which there are 100) should each take 33ms or less. So the maximum possible elapsed time would be 100 * 33ms = 3300ms = 3.30s. We’re getting a tenth of that. That’s why it says “and no timing issues”. It would report any tests that were too slow if there were any.

  4. Trackback from dahlia's me2DAY:

    홍민희의 알림…

    Full Pass of Acid3…

  5. aljoscha Says:

    Mmh, on first load I get the error “kungFuDeathGrip is null”, that vanishes upon reload.
    Seems to be https://bugs.webkit.org/show_bug.cgi?id=18144
    Alex

  6. Trackback from bomnun's me2DAY:

    봄눈의 생각…

    사파리가 가장 먼저 Acid3 100점 받았네…

  7. Trackback from meneame.net:

    Webkit, el primer motor de navegadores que pasa el test Acid3…

    El engine Webkit usado por navegadores como Safari o el nuevo Google Chrome se ha convertido en el primer motor que pasa el test Acid3 que prueba que los navegadores cumplen los estándares web, especialmente los de DOM y ECMAScript, muy usados en la W…

  8. doekman Says:

    The WebKit-r36882 on Windows doesn’t even pass. I get 98/100…

  9. wellofsouls Says:

    I thought WebKit nightlies have long fully passed Acid3 test for months already? The first being r34278 in May?

  10. flyer Says:

    revision 36882 does not pass acid3 on my mac mini. 3 test failed:
    Test 77 failed: expected ’4476′ but got ’5550′…
    Test 78 failed: expected ’90′ but got ’0′…
    Test 80 failed: kungFuDeathGrip was null
    Total elapsed time: 1.81s

  11. jimbob Says:

    OK. Now can we get it to render real-life web sites worth a damn? Thanks.

  12. Oliver Says:

    @jimbob: it does render real websites. If you have a site that webkit fails to render correctly you hsould file a bug at http://bugs.webkit.org

  13. coolfactor Says:

    Congratulations to the WebKit team. You should all feel very proud.

    @jimbob – please back up your statement with some examples. Boggles my mind why you’ve posted a negative comment on a blog for a *volunteer* team. What have you contributed lately that’s possible? So far, I’ve seen nothing.
    (sorry for sidetracking, but wow, the nerve)

  14. coolfactor Says:

    “positive”, not “possible” above.

  15. gushenry Says:

    Um, I’m not sure if this was supposed to happen, but I got that same dialogue box about a month ago…

  16. Nomis101 Says:

    OH, please help!
    I’ve downloaded the newest nightly (r36882) to take the Acid3 test, but I don’t get 100/100.
    If I take the test for the first time I get this: http://img519.imageshack.us/img519/8302/firsthp0.jpg
    And if I reload the page I get this (after every new reload): http://img218.imageshack.us/img218/8762/secondjp2.jpg
    What could be the reason for this?

  17. Nautivar Says:

    —————–
    Well not 100% (probably 99%)
    —————–
    As stated on the link (smooth animation)
    and on the wikipedia entry…
    “… and must not only be pixel-identical with the reference rendering,
    including the favicon… ”
    —————–
    I see no favicon in the rendering screenshot above!

  18. owlboy Says:

    I can only get to 69% on my G5 with a recent nightly. Here is the crash log:

    http://tinypaste.com/7c49d

  19. ajmas Says:

    I have just been testing r36882 on Windows XP. While most of the time it reaches 100/100, I am sometimes finding that it hits 99/100. Just refresh the page a few times to see. It is 99% perfect, but I wonder how significant the inconsistency is in the final 1%?

  20. dak Says:

    I’ve been able to get a full pass on hardware pretty close to the reference hardware since SF was first released. However, loading Acid 3 when its not cached leads to some speed failures. Reloading generally fixes those and gives me a time around 0.55s – 0.75s. However, going back and then clicking on the Acid 3 test link again (instead of reloading) gives me times around 0.20s to 0.35s.

    This speed increase from clicking on the link as opposed to reloading the page happens for me every single time.

  21. Maciej Stachowiak Says:

    @Nautivar check the reference rendering. It’s supposed to have a generic favicon, not a custom one. It’s a test to verify that loading favicons respects error codes, because http://acid3.acidtests.org/favicon.ico returns a valid image but with a 404 response code.

  22. Maciej Stachowiak Says:

    @ajmas Some of the tests depend on networking and so may time out on a network load. A cached run will give reliable results. The procedure I suggested (load test, close window, open new window, load test again) should always give you a cached run.

  23. Nautivar Says:

    Aha Maciej, that i didn’t catch. The wording on the linked page and wikipedia are
    a bit vague in that respect. Great, now i’ll have that drink to Webkit does Acid3, tonight :-)

  24. Trackback from Home Biss:

    WebKit fully passed Acid3…

    Yes, that was an amazing achievement alright. I guess we should pay more attention to Safari from now on but let’s not forget that Google Chrome is also a WebKit-based browser. I sure hope that Google will tweak their Chrome browser further right aft…

  25. Trackback from efectotequila.com:

    Safari es el primer navegador en pasar 100% el test del Acid3 [EN]…

    Al parecer Webkit es el primer motor de renderizado web que pasa completamente el AcidTest 3. Segun lo anuncian en el blog oficial de Webkit….

  26. ddkilzer Says:

    @ owlboy You have InputManagers installed. Please disable/remove all of them and restart Safari.

  27. owlboy Says:

    @ddkilzer Removed. Same crash at 69/100

    http://tinypaste.com/f0bff

  28. Mark Rowe Says:

    @owlboy I’m going to take a wild stab in the dark and suggest that you have the Adobe SVG plug-in installed. Check Help -> Installed Plug-Ins and see if there is any mention of SVG. If there is, please remove the plug-in that it mentions (they typically live in /Library/Internet Plug-ins/) and try again. If the problem persists, please see a doctor (or file a bug report at http://bugs.webkit.org/).

  29. btreehugger Says:

    Top-notch work, guys! Even on a T41 laptop with 512mb RAM I am getting almost every test passing fully (Windows XP SP3). Only numbers 26 and 69 sometimes take longer than anticipated, and on a cold start sometimes 00 also has trouble. However, when hitting F5 to refresh it gets 98 the second time around, failing test 77 (expected ’4776′ but got ’5550′ on getComputedTextLength) and 78 (expected ’90′ but got ’0′ on getRotationOfChar(0)).

    Keep up the stellar work!

  30. ain Says:

    Very nice indeed!

    For the comparison, Opera 9.60 Beta scores 85 in the Acid3 test.

  31. owlboy Says:

    @ Mark Rowe That was exactly it! Silly Adobe, always making clunky plugins.

  32. manufaktor Says:

    You guys are AWESOME! Now I can abandon Firefox/Firebug for development :)