Latest WebKit Builds Address Browser Scrolling Performance on Retina MacBook Pro Models

[cmChimera]

The issue with the nightlies is they're highly unreliable. Every now and then an update comes in that makes it crash all day, introduce rendering errors, etc. It's meant for testing and nothing else. Only a fraction of the changes in the nightly actually gets into release WebKit.

Ah gotcha.

I haven't noticed anything besides the scrolling, then again my previous MBP probably had some lag because it was quite old. Which instances can you detect the UI lag?

The scrolling lag is completely gone with the new Webkit build by the way.

I've noticed some lag in Mission control. It's not every time, in fact it's usually the first time I use it after sleep or a long period. Afterwards it's usually fine, but I still maintain MC is slightly slower than my old MBP. I hope it actually can be fixed with software, it would be great. Overall it's a minor issue though. Also I keep my rMBP scaled to 1680x1050, so I probably get more lag than many other users.

Yep I noticed the nightly build fixed the scroll lag, can't wait for it to be implemented into Safari.

 

[iChrist]

.

Very good.

This will be decent band-aid till rMBPs with excellent performance are released with Haswell.

 

[Andj00]

who the **** eats horse

 

[viva la ben]

who the **** eats horse

Godfrey Chang eats horses

 

[KnightWRX]

Us Chrome users will probably get this waaaay before you Safari guys.

----------

.

Very good.

This will be decent band-aid till rMBPs with excellent performance are released with Haswell.

This isn't a band-aid. The issue has never been lack of GPU/CPU power. You can scroll smoothly on much older hardware running the 30" ACD, which is the same resolution as the 13" MBP Retina. It has nothing to do with pixel pushing power.

----------

This isn't the only area where lag is found. There is general UI lag in the OS as well. Not to say that there isn't some software issues there as well, but they have a lot of evidence that GPUs just aren't up to snuff yet.

The GPUs are fine, stop repeating this non-sense. My nVidia 9400m equipped unibody MacBook could run a 30" ACD with dual-link DVI without issue, at 2560x1600. Why do you think an Intel GPU 3 generations newer couldn't do the same ?

It's obviously a bug in the software stack that has to do with the way Retina scaling (HiDPI) occurs.

 

[komodrone]

Sweet.

It's a shame so many "trusted" review sites (including The Verge) dinged the rMBP and blamed the hardware.... for what is clearly a software problem.

it's the hardware. rMBP switches over to intel graphics 4000 which can't handle the amount of work needed for smooth UI

 

[NutsNGum]

.

Very good.

This will be decent band-aid till rMBPs with excellent performance are released with Haswell.

Err.. or it's a software optimisation issue, as this clearly points out.

----------

it's the hardware. rMBP switches over to intel graphics 4000 which can't handle the amount of work needed for smooth UI

Bullcrap.

 

[2crazy]

Facebook scrolls at 0fps on my MBPr

Sorry for the trolling! I just don't get Facebook!

 

[TSloper]

So why did Apple release a product that clearly wasn't ready yet?

It clearly wasn't ready. I've had to browse using this beautiful display for months now. If only someone had told me I wouldn't be scrolling at 30 fps or better I would have never put myself through such hell. I will never buy another device again as it clearly won't be as ready as the next generation.



The hit in frame rate (hardly an issue for everyday tasks) is far outweighed by the visual quality of the display.

 

[keviig]

Thank you!

To the guys that made this and to MacRumors for posting this.... THANK YOU! Scrolling has been a nightmare on my 15" rMBP ever since i got it and this improved it by ALOT! Finally it's useable for the now-and-then facebooking

 

[Uplift]

Strange, i've never really had any issues scrolling on Safari, i've mainly been a Chrome user which runs like absolute crap on my rMBP so i've been switching back to Safari, which runs great.

 

[valexa]

This isn't the only area where lag is found. There is general UI lag in the OS as well. Not to say that there isn't some software issues there as well, but they have a lot of evidence that GPUs just aren't up to snuff yet.

I am seeing roughly the same numbers on a five year old iMac and scrolling trough google plus (about 30 fps in Chrome, about 10 fps in Safari 6.0, about 40 fps in Webkit nightly) so any performance changes should apply across the board.

I cant say I've ever noticed this myself. I'm assuming it only happens in certain circumstances?

(Using Chrome, 15" rMBP)

Chrome .. ?! hint .. hint, sure WebKit powers both Safari and Chrome but this was specific to bad Safari performance, its looks marginally acceptable in Chrome as it is.

 

[blackhand1001]

That was ages ago. Today's webkit is a much larger beast than that old KHTML. Go ask Dave Hyatt how much work he's put into Webkit as an Apple employee.

To say Apple hasn't make a significant investment in Webkit is to be either ignorant or foolish or both.

There are tons of other contributors to the project, in the large scheme of things apple has only a small role.

 

[3587]

The latest Webkit is noticeably smoother on my MBPr 15"! Way better! Scores a 388 compared to a previous 378.

 

[50548]

Thank god. It's the worst part of this expensive machine. Think I'll just install the nightly rather than wait.

----------

Now bring us a Retina CS6, upgradable RAM and we'll be cruising .

In other words, the BS argument spread by AnandTech, Ars Technica et al. that the Retina's hardware was not powerful enough to drive that great "number of pixels" (thus requiring a future GPU update, nothing less) has been buried deep in the ground...it was all a matter of software optimization.

Where are the pundits now?

 

[cmChimera]

I am seeing roughly the same numbers on a five year old iMac and scrolling trough google plus (about 30 fps in Chrome, about 10 fps in Safari 6.0, about 40 fps in Webkit nightly) so any performance changes should apply across the board.

Image




Chrome .. ?! hint .. hint, sure WebKit powers both Safari and Chrome but this was specific to bad Safari performance, its looks marginally acceptable in Chrome as it is.

I think you misread my post. I'm talking about OS X UI lag, not web browser scrolling lag.

 

[Saladinos]

Noticeable improvement when using the nightly - especially while zoomed.

Hope we get Safari 6.1 soon then. Probably along with 10.8.3.

Also, whatever happened to that password sync thing? I'm a little tired of my iPhone, Mac and iPad all requiring me to log in to websites again based on which one I signed in from last. Hopefully they could sync authentication tokens or whatnot to smooth that out.

 

[thespazz]

Holy COW. Apple needs to break this down and apply it to the OS as a whole. This looks GOOD.

 

[MacFoodPoisoner]

Great news, so, hopefully soon EVERYTHING BUT safari will lag and stutter in the rmbp... What a triumph...

This isn't a band-aid. The issue has never been lack of GPU/CPU power. You can scroll smoothly on much older hardware running the 30" ACD, which is the same resolution as the 13" MBP Retina. It has nothing to do with pixel pushing power.

That's incorrect and has been pointed out a gazilionth times here. It has everything to do with pixel pushing power, the ACD might have the same resolution with the 13" mbp, in the later the resolution is doubled and then halved (or adjusted to some %) to fit the screen. Thus it's the overhead to run 4 ACDs (no of pixel quadrupled) not one, plus the processing cost of then adjusting it to the actual screen size. When the no of pixels you render are QUADRUPLED you bet it has everything to do with pixel pushing power.

 

[KnightWRX]

I think you misread my post. I'm talking about OS X UI lag, not web browser scrolling lag.

Which in all probability is the same issue. Again folks :

30" ACD ran at 2560x1600 like 5-6 years ago. GPUs aren't the problem, it's something in the HiDPI software stack.

I personally wouldn't trade my rMBP 15" for the world.

 

[CausticPuppy]

it's the hardware. rMBP switches over to intel graphics 4000 which can't handle the amount of work needed for smooth UI

Umm, the 15" rMBP does not switch over to the discrete graphics every time you scroll down a web page.

Reviews like The Verge made comments that laggy/chopping scrolling were a hardware issue.

The article referenced in OP (did you read the article?) points to software optimization instead.

In fact, Anand's review of the 15" even stated that using the more powerful discrete chip didn't help browser scrolling much at all, because it is single-threaded-CPU dependent, rather than GPU dependent. If it's hardware issue, then why doesn't the much more powerful GPU help the browser scrolling/lagging?

 

[MacFoodPoisoner]

In other words, the BS argument spread by AnandTech, Ars Technica et al. that the Retina's hardware was not powerful enough to drive that great "number of pixels" (thus requiring a future GPU update, nothing less) has been buried deep in the ground...it was all a matter of software optimization.

Where are the pundits now?

It's not a bs argument, it's not powerful enough indeed. The ui still lags and stutters. The current hardware is just not capable enough. It's an air cpu circa rev. 1. Apple have hacked their way to resolution independence with rendering twice the resolution canvas so as to avoid what win 8 did and go into detail about interface elements. They thus offloaded all the complexity on the hardware, and the hardware isn't there yet. Simple as that. Rev. 1 retina macbook pros are a major compromise. I am still planning to go ahead and buy one, fyi, but If I do I ll be do it in fully aware that I am buying a problematic rev. 1 product, not only for the image retention issue with the lg screen, but also for the gfx barely making it there.

 

[KnightWRX]

That's incorrect and has been pointed out a gazilionth times here. It has everything to do with pixel pushing power, the ACD might have the same resolution with the 13" mbp, in the later the resolution is doubled and then halved (or adjusted to some %) to fit the screen. Thus it's the overhead to run 4 ACDs (no of pixel quadrupled) not one, plus the processing cost of then adjusting it to the actual screen size. When the no of pixels you render are QUADRUPLED you bet it has everything to do with pixel pushing power.

You're saying the same thing I am except making a few wrong assumptions. The resolution isn't doubled then halved at all in HiDPI mode, it's only downscaled after upscaling if you use the scaled modes (1680x1050 or 1920x1200). Running at 1440x900 (2880x1800) only upscales non-Retina graphics to 2880x1800 and leaves it at that. Running 1920x1200 (which I run permenantly) runs a 3840x2400 framebuffer which it then downscales to 2880x1800 for displaying on the display.

And what I've been saying all along is quite correct, the problem is not the pixel pushing ability of the GPU. Pixel fill rates, look them up. Again, if you can't understand that a 4 year old GPU, the 9400m, could power a 30" ACD and that we're now in 2012, 4 years later, with GPUs 3 generations newer, I don't know how to explain it to you.

The 9400m could push 2.32 GP/s. Again, 4 years ago. 2880x1800 at 60 fps is roughly 311 MP/s. Are you getting this ? Even if what you say is true and HiDPI was so horridly optimized as you claim (which it isn't), that's still only half of the pixel fill rate of a now 4 year old Integrated GPU.

Anyway, I own a rMBP 15" and I've hardly met any "lag" whatsoever. I guess some people are just too sensitive. Thank god Apple knows it's a software issue and is fixing it for those people.

----------

It's not a bs argument, it's not powerful enough indeed. The ui still lags and stutters. The current hardware is just not capable enough. It's an air cpu circa rev. 1. Apple have hacked their way to resolution independence with rendering twice the resolution canvas so as to avoid what win 8 did and go into detail about interface elements. They thus offloaded all the complexity on the hardware, and the hardware isn't there yet. Simple as that. Rev. 1 retina macbook pros are a major compromise. I am still planning to go ahead and buy one, fyi, but If I do I ll be do it in fully aware that I am buying a problematic rev. 1 product, not only for the image retention issue with the lg screen, but also for the gfx barely making it there.

The hardware is capable. You just don't understand GPU hardware or HiDPI to make such claims. Anand has always been full of it.

 

[iChrist]

Us Chrome users will probably get this waaaay before you Safari guys.

----------



This isn't a band-aid. The issue has never been lack of GPU/CPU power. You can scroll smoothly on much older hardware running the 30" ACD, which is the same resolution as the 13" MBP Retina. It has nothing to do with pixel pushing power.

----------



The GPUs are fine, stop repeating this non-sense. My nVidia 9400m equipped unibody MacBook could run a 30" ACD with dual-link DVI without issue, at 2560x1600. Why do you think an Intel GPU 3 generations newer couldn't do the same ?

It's obviously a bug in the software stack that has to do with the way Retina scaling (HiDPI) occurs.

Sorry my friend. The things you are comparing is like comparing apples to oranges.



----------

In other words, the BS argument spread by AnandTech, Ars Technica et al. that the Retina's hardware was not powerful enough to drive that great "number of pixels" (thus requiring a future GPU update, nothing less) has been buried deep in the ground...it was all a matter of software optimization.

Where are the pundits now?

Pundits are waiting for software and hardware to catch up and for price to come to a reasonable amount. I only mention price b.c many people consider it. I bought 7 for the office but will be 1st to admit it is definitely a revA product that will be seriously improved by Haswell.

 

[Mac Rules]

You're saying the same thing I am except making a few wrong assumptions. The resolution isn't doubled then halved at all in HiDPI mode, it's only downscaled after upscaling if you use the scaled modes (1680x1050 or 1920x1200). Running at 1440x900 (2880x1800) only upscales non-Retina graphics to 2880x1800 and leaves it at that. Running 1920x1200 (which I run permenantly) runs a 3840x2400 framebuffer which it then downscales to 2880x1800 for displaying on the display.

And what I've been saying all along is quite correct, the problem is not the pixel pushing ability of the GPU. Pixel fill rates, look them up. Again, if you can't understand that a 4 year old GPU, the 9400m, could power a 30" ACD and that we're now in 2012, 4 years later, with GPUs 3 generations newer, I don't know how to explain it to you.

The 9400m could push 2.32 GP/s. Again, 4 years ago. 2880x1800 at 60 fps is roughly 311 MP/s. Are you getting this ? Even if what you say is true and HiDPI was so horridly optimized as you claim (which it isn't), that's still only half of the pixel fill rate of a now 4 year old Integrated GPU.

Anyway, I own a rMBP 15" and I've hardly met any "lag" whatsoever. I guess some people are just too sensitive. Thank god Apple knows it's a software issue and is fixing it for those people.

Spot on.