Comments on: Safari challenges Chrome on Web app speed

Google's browser rules the roost when it comes to JavaScript speed, but Safari is the first to come anywhere close on Google's V8 benchmark.

[houltmac]

After a bit of a time installing Safari 4 is now working great here. I had real trouble installing it. It was an Input Manager, SogudiI would imagine. Now it's up and Safari140 works, but Inquisitor is broken which bums me out.

I didn't like the tabs to begin with and just saw it as a Chrome rip off, but it actually saves a ton of space. It's super fast all in and seems stable so far. Still doesn't auto-open the previous session which sucks though. The eye candy is nice and I might use the Top Sites thing - I used to code my own Web0.2 version of this back in the day before "Open folder in tabs".

Not bad so far though. Only one bug found but that was already in Safari 3 also by the look of it (check out my Flickr stream for that).

[dabepa]

Are you sure all your numbers are correct? I just ran the SunSpider test and got a score of 849.8 ms for Safari 4. That's quite a bit different from your 4,257 figure for Safari 4. Further bringing your data into question: <http://crave.cnet.co.uk/software/0,39029471,49301219,00.htm>. crave says "our benchmarks confirm Apple's Safari 4 browser, released in beta today, is the fastest browser on the planet." I ran SunSpider for Safari 4 and the latest Firefox and Safari 4 beat it by miles. I would have run it on Chrome too but I'm on OS X currently and the other numbers I've seen show Safari beating Chrome as well.

[Shankland]

Benchmarks vary widely from one machine to another. I ran the tests multiple times and used averages. Perhaps things look different if you compare FIrefox vs Safari on Mac OS X vs on Windows as I did. The numbers certainly are different if you have a more or less powerful system overall. I'll run some tests on a quad-core 64-bit Windows Vista system too and see what comes up.

[Goodbye Helicopter]

he probably ran these on windows

[jwhitehu]

I tried to load Safari 4 on a MBP and Imac with the same result which ended in no installation at all and several attempts to relaunch. I am running the latest version of Leopard on both. Cant figure out why the installations keep crashing and wont open at all. Guess I'm back to 3.2 until 4.0 comes out in a proven install

[Gnarlodious]

Java on Mac runs faster than Windows.

[seven7dust]

why care about numbers so much !
the UI and real world performance r more important than speed !

I see these types of comparisons even on mobile browsers and phones
which is even more ridiculous !
cause UI rules the roost in a browser, especially in a phone !
Becnhmarks r secondary IMO !

as far as the new interface goes
I'm disappointed that they decided to remove the blue completion bar
the rest of the UI quirks r a good improvemnt , I like the new tab bar especially
but I expected more on the search front !

[JJesper]

Isn't it time for us to start benchmarking actual applications rather than synthetic benchmarks? As the differences between V8 and SunSpider results show, synthetic benchmarks aren't in any way objective.

In CPU benchmarks sure there is a place for synthetic benchmarks, but nobody would write an article on a CPU without running benchmarks of real applications.

I am also afraid that the way the current press coverage focuses SunSpider benchmarks in effect forces browser developers to optimize for benchmarks rather than for actual application speed.

I would love to see CNET take the lead here and start benchmarking the applications we actually use.

[Alan4ik]

I think results are depend on the OS, hardware and internet connection therefore results could vary. I ran the SunSpider test on iMac Intel Duo Core 1.83MHz, 2G of RAM Mac OSX Leopard with the broadband connection of up to 10Mb.
FireFox 3.0.6 4481.8 ms
Ferfox 3.1 beta 2 2114.4 ms
Safari 4 1172.6 ms

The test took place in the UK between 0.00 - 0.05 a.m

[Angmarr]

umm.. Just use Firefox

[seven7dust]

no safari is faster and less of a resource hog than firefox for me !
the days of Firefox being the only alternative to IE r over
Chrome and safari own Firefox in every way possible !
Extensions r only gimicks ,if n e thing they r more of hassle than a advantage !

[Mark_Anderson]

Adblock, noscript and Downloadhelper are 'gimmicks'?

Oh my!

[pithenumber]

@77dust
FireFox is far superior to Safari
I have tried Safari on the Hackintosh and on Windows, Firefox is better

Extensions = gimmicks? NO!
I love my extensions, adblock, noscript, firebug, greasemonkey
they are 100% awesome and 0% hassle

[seven7dust]

@pithnumber, @Mark_Andrerson

all that is possible on both safari and opera without extensions
thats the problem with FF, it requires extensions for every single thing
it adds more bloat to a already bloated browser !
I agree that safari 4 is not good in windows
but safari 3.1.1 is way better than firefox

[ballmerisanape]

This thing rocks so far. I have it on a 12 inch powerbook 1 ghz G4 with 1.25 Gb ram and OS 10.5.6.... and I have Safari AdBlock installed.

So far very zippy. It took about 30 seconds for all of my favorites to display their icons... but after that... smooth as silk....

[ausernamenoonehaschosen]

I installed it last night on my 867 MHz 12" PB, and found it to be very nice as well. Something else, before in Safari 3 hulu flash videos would jump a bit, now they are smooth. There are a lot of additions to this version, and so far I am very impressed.

[Clarious]

I see it is interesting that Opera 10 alpha have pretty high score, even higher than firefox 3.1b in Google V8 benchmark. Maybe because you use the raw version, tracemonkey (or whatever its name is) wasn't turned on?

And you should have use the same colors for browsers in two benchmark, it look a little bit confusing at first :)

[Shankland]

Sorry about the colors--you're right. The tool I was using, Google Docs, doesn't permit that level of control.

Firefox 3.1b2 does have TraceMonkey enabled by default.

[digiworkz]

I just ran the sunspider test on a 2.33 MBP with 4gb ram using Mac OS 10.5.6 with a 7mb internet connection.

The test scored 3322.6ms with +/- 0.7%.

I think he was using the wrong version of Safari to achieve the scores above.

[Shankland]

Results vary from one computer to the next. Also, I wouldn't be surprised if Safari has a relative advantage on Mac OS X--Apple could optimize more there and Firefox could optimize more on Windows. I don't have any hard and fast data about that, though.

[kcotham]

As far as UI changes are concerned, Safari 4 is a major step backwards. ALL Mac OS Applications have a static top frame for the window. The tabs only took up 5 mm of screen space on my Early 2008 Mac Book. And since this is the smallest, lowest resolution screen on a Mac that is "modern" then it should be worst on it. Five millimetres is not much at all. And it definitely isn't worth ruining the entire UI over.

The Reload button has been put in the address bar. WHY!? That is the dumbest thing I've ever seen. All in all, this is a very poor excuse for UI "improvements". It seems that Apple has lost their objectivity and their own direction. Ever since they've made software like Safari available for Windows they've been too concerned with courting Windows users. Go back to your roots Apple. Forget what Google or Microsoft are doing. Don't try to appease Microsoft Windows users. Follow your own Human Interface Guidelines for crying out loud!

[Bizu1969]

OMG!!!!!! APPLE YOU OUT DID YOURSELF!!! this new Safari is AMAZING!!!!! DO NOT TAKE IT AWAY!!!!!!
HOLY &@^#%@&^#% I love it.. i work on my 32" monitor and i am sitting here drooling.. AND it's fast... who cares about comparing speed with Chrome...APPLE.. remember this...your look no matter what you do is ALWAYS cutting edge.. and everyone else is chasing behind you trying DESPERATELY to be like or as good as you. SAFARI 4 FTW!!!!!!!!! :) :) :)

[pithenumber]

did you write this when you were drunk or are you just the biggest !@#$%^& fanboy out there

[generallysceptic]

Hmm the first thing I did after installing Safari 4 was browse to flickr and it immediately crashed... I just love how people all over the net are singing praise for Safari 4 when its just really a beta and apparently even more unstable than Minefield, which was about as fast in most benchmarks anyway. Chrome is also about as fast and probably 10x as stable, since its been tested extensively (Im not a fan though). As said above, benchmarks dont really mean a lot and results will always vary greatly, so the only thing that can really be a deciding factor is features, extensions and above all stability.. and Im not seeing that Safari 4 is all that stable at the moment.

[G0reTan]

its a beta u idiot.....do you know what beta means? you should expect some glitches and errors. geez. people these days.

[pithenumber]

@G0reTan
We have come to expect more out of a Beta ever since a near perfectly stable Windows 7 BETA came out

[heatwaveo8]

After installing it, it took about 10 attempts for it to finally open properly, but its been working well since. :)

The results are going to be different for everyone depending on your OS and your connection speed and type. For myself, being a Mac user on Broadband, I've found the new Safari version to be the fastest one for my computer now, given that Chrome isn't available for Mac yet. It's definately faster overall than Firefox, which is what I had been using before since it was faster than Safari 3. (Opera and Firefox are about the same speed for me, but Firefox has had more features.) I'm willing to deal with the bugs for a little while if any appear, though, because the thing that matters most to me is the navigation speed.

[doctorSpoc]

yes the absolute value of the numbers will be different but the ranking should be the same for a platform... and this guys rankings are different than everyone else's on the planet. others have done tests on windows including myself an not got these results.

RAM might be a factor if one browser depends on it more than another but.. he tested with 3Gig so there should be no problem there... there can be no other conclusion that he's either just made up the numbers or is wildly incompetent.

[jordanaustino]

Um... I get consistant 950's on the sunspider speed test with chrome 2.0.164, nothing comes anywhere near its speed and safari doesnt have a ctrl -t new tab function, makes my webbrowsing slow

[doctorSpoc]

as i said.. guys run the test yourself if you want REAL results... don't believe these bogus results...

Windows XP Prof, SP2, Intel Xenon 3GHz, 3Gig RAM...

here's what i got..

i did Acid Tests on..

Safari 4 beta 100/100
FireFox 3.1 B2 93/100
Chrome 1.01.154.48 78/100
FireFox 3.0.5 70/100
IE 7.05730.13 12/100 (this was pretty funny... was a jumbled mess piled in the upper right corner ***?)


SunSpider JavaScript benchmark... can also use V8 test...

Safari 4 beta 654ms
Chrome 1.01.154.48 819ms (1.25x safari)
FireFox 3.1 B2 1054ms (1.6x safari)
FireFox 3.0.5 2234ms (3.4x safari)
IE 7.05730.13 20340ms (31.1x safari)

[doctorSpoc]

obviously haven't tried safari since it does indeed have ctrl-t to open a new tab... where did you get that from?

[Llib Setag]

CNET UK BENCHMARK TESTS are completely the OPPOSITE of what you are saying...WHO IS LYING & HOW MUCH ARE YOU GETTING PAID TO LIE?

http://crave.cnet.co.uk/software/0,39029471,49301219,00.htm

[doctorSpoc]

sorry man, but you numbers a completely boggus... i ran sunspider on a 3GHz Xenon with 3GB RAM running XP and Safari ran 1.2x faster than Chrome.. take a look at the results from adrian kingsley on ZDNet.. his numbers jibe pretty much exactly with what i got...

you guys should run sunSpider yourself because the number presented here don't match what i got or any number i've seen from any other review i've seen online thus far... the only explaination i can see is that the reviewer must have been running some photoshop filter or encoding video while running the test on Safari.

don't be snowed guys these results are completely bogus

[ukdude]

So this is not comparing Safari 4 which scores 100% on Acid3 with 78%-scoring Chrome (which is slower on SunSpider), but with an experimental Chrome build. A bit like comparing a street legal production car with a specially modified hotrod!

They are both fine browsers, of course, but right now Safari is a far more robust choice than Chrome.