New Peacekeeper Beta Version Launches
Give it a go.
Good news for all you benchmark geeks, Futuremark, the world’s leading benchmark maker, has released a new version of its highly popular web browsers testing tool for you to play with.
Peacekeeper 2 implements the latest HTML5 standards and features and was designed to work with web browsers running on any operating system and pretty much on every device (including tablets and smartphones).
While still in Beta, Peacekeeper developers are expecting to receive some constructive feedback and/or suggestions in the comment section below.
So, launch it now and let them know what you think.
About (Author Profile)
Vygantas is a former web designer whose projects are used by companies such as AMD, NVIDIA and departed Westood Studios. Being passionate about software, Vygantas began his journalism career back in 2007 when he founded FavBrowser.com. Having said that, he is also an adrenaline junkie who enjoys good books, fitness activities and Forex trading.
Subscribe
If you enjoyed this article, subscribe to receive more just like it.
nice to see they finally updated it… but now it doesnt work
Test timeout exceeded while running test workerContrast01. in FF 3.6.23
Get rid of the LSD in your system and try again s
same error in FF 7.0
Test timeout exceeded while running test workerContrast01.
i should be able to skip the HTML5 feature tests since they dont count towards the score but currently the only option is to start again (which ends in the same error)
Hi,
I’m the guy in charge of Peacekeeper at Futuremark. I can verify that the beta does have issues with FF3 on the worker test – for some reason, FF3 claims to have the ability to run the test but cannot complete it. It does, however, work on FF5, FF6 and FF7.
Opera manages to do the first one but gets stuck on the second only managing 1 out of 3 images
Test timeout exceeded while running test workerContrast02.
i dont think this cpu test works very well on 1ghz machines unless they have insanely fast js speed like chrome and only being able to test one browser doesn’t make it very useful…
Hrm. Can you tell me what kind of a system you have? I was worried about the worker test being too heavy for low-end systems, but I have successfully (and repeatably) ran it on a Win7 tablet. It’s slow but it does finish.
The timeout is there to catch situations where the browser is likely stuck and will never finish. If necessary, I can make the timeout period be a bit longer to allow slower systems to complete the test.
its a 10 yr old AMD single core machine
I’ll see if there is anything we can do about this. Unfortunately when doing the tests we need to make sure that they are sufficiently taxing for high-end systems as well, thus very slow systems may be unable to complete the workloads in the time.
updated test is working fine now in firefox :)
at least it did in 3.6.23 because it skipped the test
workerContrast02. still times out in FF8
sore wa nanda?
Opera ran the old test fastest on my PC with Peacekeeper 2 Chromium is the champ but the test does seem buggy Chromium got a yes for supporting 2 tests while only displaying a black box.
Visually I would have picked Firefox as the winner instead of the third place it got .
http://peacekeeper.futuremark.com/results?key=4Cpn&resultId=1840988
Recently it was in news that the devs at Opera have optimized the browser especially for the Peacekeeper test mechanism. that’s why you see Firefox Nightlies and Chrome dev scoring much less than Opera in Peacekeeper bench.
I was wondering how did Opera manage to take a leap above Chrome and Firefox despite not having hardware accelerated Canvas and graphics. Either it’s a greatly coded engine or something else, no one knows. Peacekeeper isn’t a very reliable test, in my opinion.
BTW, I’m an full time Opera and Firefox user, to clear any doubts.
Any links about Opera optimizing for Peacekeeper, haven’t seen anything personally.
What leap beyond Chrome ? in the old Peacekeeper test Hardware acceleration was not used.
In the new Peacekeeper Beta test Chrome is fastest for now.
Different benchmarks test different things so just because a browser is fastest in Sunspider it does not follow that it will be fastest in every benchmark.
I too remember this optimization claim.
LMGTFY: http://dev.opera.com/forums/topic/398731
http://my.opera.com/chooseopera/blog/a-new-era-of-browser-speed
Though they are nothing recent
I’d also love to see information on anyone specifically optimizing for Peacekeeper. Can you provide a link to this bit of news?
I’d also love to see information on where peacekeeper diverges completely from other benchmarks. We do differ a bit as we do have different tests, but in my experience there has not been a vast difference.
V8 is not reliable because it was designed for Chrome. It drops all the stuff Chrome is slow at for some reason.
I saw a similar issue during testing where the videos plays but does not show on the page (during testing we had sound in the clip so we could tell). I thought we had ironed it out but apparently it still needs looking into.
Test works fine now: http://peacekeeper.futuremark.com/results?key=4Ctt&resultId=1841382
why is chrome considered king? in every test i’ve ever done, it has been almost exactly level with opera. chrome has some better text and array handling, but opera wins in every other aspect. if i had to choose between javascript or rendering performance, i know what i’d choose.
People run the test & so far Chrome has been fastest
that is because it puts too much weight on the less important tests.
The score is calculated as the geometric mean across all suite scores. The suite scores themselves are the geometric mean of the tests within them. As such, there are no specific weights that make tests of a certain suite more important than others.
Having no specific weights means all tests are valued equally – something he is saying they shouldn’t be.
Adding in specific weights would very definitely make the benchmark not be as it should be. I would not want determine how much more important, say, arrays are to strings.
In terms of canvas drawing, Opera is definitely the slowest browser at all.
Are we even using the same browser? Or is your assessment based on 5 minutes of trying Opera?
Have you run the Peacekeeper Beta test ?
On my PC HTML5 Cavas Firefox is fastest followed by Opera then Chrome in third place.
In the original Peacekeeper test Opera was much faster in the Complex graphics section which uses Canvas
http://clients.futuremark.com/peacekeeper/results.action?key=6QYc
The canvas test(s) are copied straight from the original peacekeeper, so I’d be very curious to see if there is a situation where browsers perform differently in them between the old and the new. Can you provide result ids?
Sure here u go
old http://clients.futuremark.com/peacekeeper/results.action?key=6QYc
new http://peacekeeper.futuremark.com/results?key=4Cpn&resultId=1840988
My scores were
Chrome16 5046Opera12 4006Opera12 3961Firefox10 2809
One very noticeable difference between the old & new Canvas tests there is no Space star wars video.
Also in the ripple tests visually I would have said Opera was a bit faster than Firefox
Opera scored
HTML5 Canvas 32.45experimentalRipple01 39.90 fps experimentalRipple02 26.39 fps
Firefox scored
HTML5 Canvas 41.73experimentalRipple01 61.37 fps experimentalRipple02 28.38 fps
Hrm there definitely is a difference in the score. We did change the timing method to use animationframe rather that settimeout, but the effect should be fairly small.
The suite should imho contain more than just old tests with new graphics – there is plenty of new HTML5 features to be tested, just look at IE Test Drive page :)
For example adding some jQuery selector/replacement benchmark will be really cool :)
Also – the only login possibility using facebook, wtf? Add it at least optional, or better use OpenID login :)
Facebook was just the first one we did, we will add others ;)
I agree a jQuery selector test would be fairly interesting. I’ll see what I can do.
Are you part of the Peacekeeper team? If so, please keep standards tests and performance tests separate. How much html5 you support is not relevant when someone wants to see how fast a browser is, and vice versa. It would be a shame to see a performance benchmark littered by tests for standards that don’t tell you anything about performance.
I am, yes. Our intent is not to start testing conformance at large, but we feel that testing the most interesting features does have merit. I believe there are a fair amount of users who have never seen WebGL before.
If we could, we would have performance metrics measured from all tests. Right now the only one that does not have it at all are the video tests, as the tag does not offer any real information on how well the video within is playing. Should this change we’ll add it in.
In my pc opera scores similar to chrome without doing all the tests. Opera only could do 4 of the 6 tests, and chrome have done 5 of the 6 tests. (in html5 capabilities).
Opera FTW.
Note that the html5 capabilities suite is not calculated into the overall score :)