Google Chrome Causing Freezing and Crashing on New Mac Notebooks

[faroZ06]

Geez, because it's so hard to just disable the plugin :

View attachment 345741

Or make it on demand :

View attachment 345742

Yep, let's just call Chrome wasteful instead of useful. Safari running Flash has the same problem.

Did you think I was comparing Chrome running Flash to Safari not running Flash? I was comparing the Flash component in Chrome to the one in other browsers, and then I compared Chrome and Safari idling (no Flash).

----------

Not in my experience. On my late 2006 iMac 24' running Snow Leopard, Safari and Firefox are slower than a slow motion replay of a snail race during the Siberian mid-winter championships.
Chrome just works.

Something must be wrong. I actually have the same iMac as you but a different OS (Leopard).

 

[wordoflife]

May I ask why? What are you really missing in Firefox 13?

Anyways, Firefox has been incredibly stable for me these past rapid releases. Much more improvement has been gone into memory management so much so that Firefox uses MUCH less memory than Chrome and it seems so much faster as a result as well. Give Firefox another try. I'm using the bleeding edge version (Firefox 16 UX) and for a pre-alpha version, it is super fast, and most surprisingly, it's so stable. Plus it has Lion goodies like the Lion scrollbars and Fullscreen.

Opera is also another solid offering. I tried Opera 12 and the only thing preventing me really from switching to Opera is the lack of any alternatives to Firefox's amazing array of addons.

I don't really see the appeal of Chrome. I keep it if I want to watch flash videos since I don't like to install Flash on my system and to me, Chrome is much less configurable than either Firefox or Opera and I really don't like some of it's weird quirks.

It's not really that. Firefox used to be the best thing back in the day and that's why I liked using it. It was better than anything else.

But version 3.0 and onwards has just been the same, and I'm really starting to appreciate features in other browsers such as top sites (i know the latest version of FF has it, but they got around to it way longer after everyone else), reading list (safari), and the ability to sync bookmarks (chrome, safari).

It's also the small things. When you have a lot of tabs open, the actual tab gets smaller. Subseqeuntly, the the 'X" button to close the tab changes position. In Chrome, when you have a lot of tabs open and you want to close a lot of them at the same time, you keep pressing the "x" on the tab, but the tab size won't resize until you move your mouse away. On FF, you have to keep moving the mouse to close the tab.

It's hard to explain.

 

[Watabou]

It's not really that. Firefox used to be the best thing back in the day and that's why I liked using it. It was better than anything else.

But version 3.0 and onwards has just been the same, and I'm really starting to appreciate features in other browsers such as top sites (i know the latest version of FF has it, but they got around to it way longer after everyone else), reading list (safari), and the ability to sync bookmarks (chrome, safari).

It's also the small things. When you have a lot of tabs open, the actual tab gets smaller. Subseqeuntly, the the 'X" button to close the tab changes position. In Chrome, when you have a lot of tabs open and you want to close a lot of them at the same time, you keep pressing the "x" on the tab, but the tab size won't resize until you move your mouse away. On FF, you have to keep moving the mouse to close the tab.

It's hard to explain.

I can respect that. We have our own preferences.

I can't help but give a little tip though, for you or anybody else: there is a popular addon for Firefox called TabMixPlus that alleviates a lot of tab management problems and you can specify that you always want the x to show and you can even specify where you want it to show (left, right etc).

Ultimately, browsers come down to what we like to use though. I really like Firefox's customizability, security with noscript and adblock(which is better than Chrome's adblock) and the tab groups feature which is really like expose for tabs.

 

[KnightWRX]

Did you think I was comparing Chrome running Flash to Safari not running Flash? I was comparing the Flash component in Chrome to the one in other browsers, and then I compared Chrome and Safari idling (no Flash).

Then there's something wrong with your computer. Chrome does not misbehave on mine moreso than Safari. It doesn't drain the battery more than Safari. It doesn't use idle CPU cycles doing nothing more than Safari.

 

[wordoflife]

I can respect that. We have our own preferences.

I can't help but give a little tip though, for you or anybody else: there is a popular addon for Firefox called TabMixPlus that alleviates a lot of tab management problems and you can specify that you always want the x to show and you can even specify where you want it to show (left, right etc).

Ultimately, browsers come down to what we like to use though. I really like Firefox's customizability, security with noscript and adblock(which is better than Chrome's adblock) and the tab groups feature which is really like expose for tabs.

I totally agree with what you say about Firefox regarding security options and customizability. Don't get me wrong ... I still use Firefox daily. I use it for checking email, school related websites and my bank account. For my more engaging activities - "social/fun" stuff, like MR, Youtube, fb, image-board websites (hehe),games, etc ... I use Chrome. Plus I don't want my silly stuff to get mixed in with my important stuff so that's why I use different browsers. I know I can just use different profiles on Chrome but mehh. After all, i still like FF.

BTW thanks for the tip. You understood exactly what I was trying to explain, lol.

 

[CountSessine]

I've had plenty of crashes of the GPU drivers under windows 7 and they have not resulted in blue screens, the only way I blue screen my PC is if I overclock it to far.

That's odd. Assuming that you're not talking about some user-space driver component crashing, I would guess that you're seeing the Windows equivalent of a kernel 'oops'. What does that look like in Windows?

Weather its window or OS X, no program should be crashing the OS. Especially under OS X as apple controls the drivers, in a closed environment.

Yes - in the Mac world, the buck stops at Apple. Whether the bug is deep in NVidia's core driver logic, or some Apple-written interfacing logic, it's Apple's responsibility to get this sorted.

 

[YGAB]

i'd like to know if Google developers intentionally wrote a code that f**ks with Macs, or vice versa.

 

[spazzcat]

errr are you saying Apple wouldn't beta test a laptop with ground breaking new graphics with one of the worlds biggest Internet Browsers? or games or video editing software..?

na you right why should they, they probably just package it up and only run Apple software saying it'll be fine as it goes out the door

Apple isn't going to take the time to test third party software, that isn't their job. That is Googles job once the new hardware come out. Apple will test their stuff and that is it. And even if they did test it, they aren't going to call Google and say hey we have this new system coming out and you need to fix this this and this.

 

[KnightWRX]

i'd like to know if Google developers intentionally wrote a code that f**ks with Macs, or vice versa.

If your userspace code can f**ks with Macs' kernel space code, it's the kernel that's at fault, not the Google or any other 3rd party developers.

----------

Apple isn't going to take the time to test third party software, that isn't their job. That is Googles job once the new hardware come out. Apple will test their stuff and that is it. And even if they did test it, they aren't going to call Google and say hey we have this new system coming out and you need to fix this this and this.

Why would they call Google to fix a kernel bug in the GPU driver stack ?

Google's bug in their Chrome unearthed a Kernel bug. That is why the system is panicking. If the Kernel didn't have the bug, Chrome would just leak memory and eventually crash on its own, not with the entire system.

In a modern OS (or heck, just in Unix in general since well Unix has been around), userspace code can't take down the system unless there's a flaw in the system.

 

[AidenShaw]

That's odd. Assuming that you're not talking about some user-space driver component crashing, I would guess that you're seeing the Windows equivalent of a kernel 'oops'. What does that look like in Windows?

It looks like the screen freezes for a few seconds, then flashes to black and refreshes, along with a popup message that the display driver has been restarted.

http://msdn.microsoft.com/en-us/library/aa480220.aspx

At a technical level, WDDM display drivers have two components, a kernel mode driver (KMD) that is very streamlined, and a user-mode driver that does most of the intense computations. With this model, most of the code is moved out of kernel mode. That is, the kernel mode piece is now solely responsible for lower-level functionality and the user mode piece takes on heavier functionality such as facilitating the translation from higher-level API constructs to direct GPU commands while maintaining application compatibility. This greatly reduces the chance of a fatal blue screen and most graphics driver-related problems result in at worst one application being affected.

WDDM also provides fault-tolerance against display driver hangs. This enables Windows Vista to detect system hangs and restart the display driver again without the need of a system reboot.

Additionally, display drivers in Windows Vista have been significantly simplified by eliminating the need to include code for the support of various device driver interfaces introduced over many years. Thus, Windows Vista implements only a single interface while ensuring that all the older drivers are recognized and function optimally.

Note that Windows can upgrade, unload, restart or reload the kernel portion of the driver without needing a reboot. That's what the "fault-tolerance" statement means - hung kernel drivers can be restarted.

 

[subsonix]

Note that Windows can upgrade, unload, restart or reload the kernel portion of the driver without needing a reboot.

The same is true on OS X with kexts.

 

[AidenShaw]

The same is true on OS X with kexts.

Thanks for the additional info - I didn't mean to knock OSX on this point, but wanted to mention that Windows can restart and reload kernel components without needing a reboot.

(It seems that many here on the forum still assume that Windows 7 works just like Windows 95 .)

 

[faroZ06]

Then there's something wrong with your computer. Chrome does not misbehave on mine moreso than Safari. It doesn't drain the battery more than Safari. It doesn't use idle CPU cycles doing nothing more than Safari.

Well by "idle" I meant on a webpage, not just with 0 windows open. Flash in Chrome seems to just use more CPU but not a crazy amount. Also, it runs 3 or so processes.

----------

i'd like to know if Google developers intentionally wrote a code that f**ks with Macs, or vice versa.

Something is wrong with the OS or the hardware if an application is causing system-wide crashing. I blame Google if a mistake on their part triggered this, but either Apple, Intel, or both are at fault for allowing it to happen.

 

[AidenShaw]

that's a *good* thing

Well by "idle" I meant on a webpage, not just with 0 windows open. Flash in Chrome seems to just use more CPU but not a crazy amount. Also, it runs 3 or so processes.

That's actually a good thing - those "extra" processes are most likely the sandboxes for protection.

And, a few extra processes mean nothing on modern systems - my Win7 workstation right now has 133 processes and 1813 threads, and I'm not doing much of anything.

Note also that many tools don't make it easy to determine shared resources - so 10 processes each using 500 MiB of RAM might only be using a total of 550 MiB of RAM if most of it is shared.

This is especially true of "process per tab" browsers, since only one in-memory copy of the browser executable code is needed, but tools may show the total private+shared for each process and mislead one into thinking that the memory usage is far greater than it actually is.

Something is wrong with the OS or the hardware if an application is causing system-wide crashing.

I blame Google if a mistake on their part triggered this, but either Apple, Intel, or both are at fault for allowing it to happen.

Your first and second sentences contradict one another.

Userland code should *never* be able to panic the OS - by definition that's a defect in the OS. It's Apple's fault, even if Chrome or Facebook or Adobe or Microsoft software happens to set up the conditions that cause Apple OSX to panic.

And by standard OEM contracts, it is Apple's fault even if the root cause is an Intel defect (hardware or reference software).

Again, Apple users are the beta testers because of Apple's passion for secrecy. Don't you think that had a significant number of "real users" been beta testers for these systems that someone would have noticed that using one of the top three browsers caused the system to stumble and crash?

 

[faroZ06]

That's actually a good thing - those "extra" processes are most likely the sandboxes for protection.

And, a few extra processes mean nothing on modern systems - my Win7 workstation right now has 133 processes and 1813 threads, and I'm not doing much of anything.

Note also that many tools don't make it easy to determine shared resources - so 10 processes each using 500 MiB of RAM might only be using a total of 550 MiB of RAM if most of it is shared.

This is especially true of "process per tab" browsers, since only one in-memory copy of the browser executable code is needed, but tools may show the total private+shared for each process and mislead one into thinking that the memory usage is far greater than it actually is.



Your first and second sentences contradict one another.

Userland code should *never* be able to panic the OS - by definition that's a defect in the OS. It's Apple's fault, even if Chrome or Facebook or Adobe or Microsoft software happens to set up the conditions that cause Apple OSX to panic.

And by standard OEM contracts, it is Apple's fault even if the root cause is an Intel defect (hardware or reference software).

Again, Apple users are the beta testers because of Apple's passion for secrecy. Don't you think that had a significant number of "real users" been beta testers for these systems that someone would have noticed that using one of the top three browsers caused the system to stumble and crash?

The extra processes use up RAM and CPU though. I'm not saying it's a bad thing to have them, but they use more resources combined than Safari does.

Also, I know it's Apple's fault if a user app crashes the OS, but I have to put some blame on Google for making a browser that does that, just like you'd have to blame the maker of a virus if the virus messes up Windows even though Windows is what let it happen. There are a lot of apps, and Chrome is the only thing that does this.

 

[Renzatic]

Also, I know it's technically Apple's fault if a user app crashes the OS, but I have to put some blame on Google for making a browser that does that, just like you'd have to blame the maker of a virus if the virus messes up Windows even though Windows is what let it happen.

That's somewhat similar to saying "I know it's technically the manufacturer's fault for making a bucket with a holes in it, but you have to at least put some blame on Johnny for trying to fill it with water".

It's a bug that needs to be fixed. Plain and simple as that.

 

[AidenShaw]

The extra processes use up RAM and CPU though. I'm not saying it's a bad thing to have them, but they use more resources combined than Safari does.

I actually said the opposite - a process that is mostly shared memory uses very little additional RAM, and an idle process uses no CPU.

Understand what your system tools are telling you.

Also, I know it's technically Apple's fault if a user app crashes the OS

That's the crux of the matter - it's Apple's fault. Period. Plain and simple. No need for "wishy-washy" words like "technically".

It's a flaw in Apple OSx. Period. Plain and simple.

 

[Fishcake21]

This is definitely an OSX problem for a few reasons.

Since last month, Chrome has been using integrated graphics option on macbook pros that carries a dedicated card so it won't be using extra battery power and issues with switching gpus.

Chrome probably takes advantage of that integrated hd3000 to increase the smoothness, etc. using GPU acceleration, I guess they do this by an API level based on the bug report.

Now pitting in with a new type of integrated graphics (because it occurs with newer macbook pros using hd 4000), there are issues because of hardware calls from Chrome that requires use of the GPU acceleration and its a mismatch.

Other OSes I guess have their own issues or even won't allow it to happen, but OSX just panics if it deems illegal to perform.

This is not the first time it happens with chrome but various of software whenever they are calling for APIS or certain use of the hardware and you have revisions changes from hardware to kernel extensions.

It is clearly apple's own fault at this. Much like the issue with Steam games needed full resources...

And its a shame too because Chrome is very smoother with such GPU accelerations.

 

[nonamelive]

Chrome sucks on Mac but rocks on Windows.

 

[AidenShaw]

...in other words...

This is definitely an OSX problem for a few reasons.

Since last month, Chrome has been using integrated graphics option on macbook pros that carries a dedicated card so it won't be using extra battery power and issues with switching gpus.

Chrome probably takes advantage of that integrated hd3000 to increase the smoothness, etc. using GPU acceleration, I guess they do this by an API level based on the bug report.

Now pitting in with a new type of integrated graphics (because it occurs with newer macbook pros using hd 4000), there are issues because of hardware calls from Chrome that requires use of the GPU acceleration and its a mismatch.

In other words, Apple screwed up the APIs that expose GPU acceleration for the HD4000.

That's what happens when you don't beta test new systems.

Love the Apple fans trying to spin this as some nefarious Google plot, when it's simply Apple's "quality assurance" group having another epic fail.

 

[faroZ06]

That's somewhat similar to saying "I know it's technically the manufacturer's fault for making a bucket with a holes in it, but you have to at least put some blame on Johnny for trying to fill it with water".

It's a bug that needs to be fixed. Plain and simple as that.

You do have to blame Johnny for filling a bucket with holes in it with water... If I know that doing something will mess up my computer, I'm not going to do it. Apple allows its users to do sudo rm -rf /* of course.

----------

That's the crux of the matter - it's Apple's fault. Period. Plain and simple. No need for "wishy-washy" words like "technically".

It's a flaw in Apple OSx. Period. Plain and simple.

Actually, I was wrong. It's not technically Apple's fault. Both Apple and Google did something wrong, and Chrome is the only thing that I've ever seen doing this. It's also a bit of a hog.

 

[Oletros]

Actually, I was wrong. It's not technically Apple's fault. Both Apple and Google did something wrong, and Chrome is the only thing that I've ever seen doing this. It's also a bit of a hog.

In Apple forums there are reports of the same KP when using different programs so technically it is Apple's fault

 

[KnightWRX]

double.

 

[KnightWRX]

Well by "idle" I meant on a webpage, not just with 0 windows open. Flash in Chrome seems to just use more CPU but not a crazy amount. Also, it runs 3 or so processes.

Processes don't really result in CPU overhead. 3 processes vs 1 process is a bunch of kernel entries for the processes and that's it. For the scheduler, it creates what is called "a drop in the ocean" type of impact.

And now we're back to Flash. So back to my original comment. Disable or make Flash on demand and Chrome won't use more CPU. That was my whole point when I replied to you the first time, you said "but it's not about Flash!" then you came back and say it is.

Which is it ? Safari vs Chrome idling on a Webpage. It is about Flash or not ? Because just siting there, Safari vs Chrome, both are equally draining the battery and equaling using the CPU. Don't change the variables on me everytime you post, that's just insane and can't ever lead to proper discussion.

Something is wrong with the OS or the hardware if an application is causing system-wide crashing. I blame Google if a mistake on their part triggered this, but either Apple, Intel, or both are at fault for allowing it to happen.

Blaming Google for what ? Bugs in software happen all the time, Apple's own is riddled with it, hence all the patches they put out. Software on the scale both Google and Apple write will always have tons of bugs that are triggered by edge conditions not trapped by unit tests in Q&A. There's just no way to account for all the variables all the time. Sometimes, a condition may never even be triggered for the entire life of the software.

Both Apple and Google did something wrong, and Chrome is the only thing that I've ever seen doing this. It's also a bit of a hog.

You keep saying Chrome is a hog, but everyone that switched to Chrome did so because of the contrary : it's blazingly fast and light. What do you feel Chrome hogs exactly ?

Looking at Chrome's task manager (yes, Chrome provides its own) you can see the break down :



As you can see, very very little CPU is being used. Memory is also quite low on the browser itself, it's the tabs which contain heavy Web 2.0 code (Facebook/Gmail) that are more memory heavy since these tend to dynamically load and reload a ton of stuff in the background, but then again, total seems to be low for a modern browser (I've seen Safari use Gigabytes of memory).

Also notice how the Flash plugin is missing. It is enabled though. Chrome only loads it if a web page asks for it. So just by virtue of Flash being there, it's not using any ressources at all. Making it completely on-demand with FlashBlock prevents it from loading unless you want to on top. So you can't even use that to say "Chrome is a hog".

Again, the more you talk about this, the more I'm thinking there's either something very wrong with your computer, or you just don't understand modern OSes and how they work exactly. Hence I'm left puzzled by how you think you are tooled to even be making the claims you are making.

 

[faroZ06]

Processes don't really result in CPU overhead. 3 processes vs 1 process is a bunch of kernel entries for the processes and that's it. For the scheduler, it creates what is called "a drop in the ocean" type of impact.

And now we're back to Flash. So back to my original comment. Disable or make Flash on demand and Chrome won't use more CPU. That was my whole point when I replied to you the first time, you said "but it's not about Flash!" then you came back and say it is.

Which is it ? Safari vs Chrome idling on a Webpage. It is about Flash or not ? Because just siting there, Safari vs Chrome, both are equally draining the battery and equaling using the CPU. Don't change the variables on me everytime you post, that's just insane and can't ever lead to proper discussion.



Blaming Google for what ? Bugs in software happen all the time, Apple's own is riddled with it, hence all the patches they put out. Software on the scale both Google and Apple write will always have tons of bugs that are triggered by edge conditions not trapped by unit tests in Q&A. There's just no way to account for all the variables all the time. Sometimes, a condition may never even be triggered for the entire life of the software.



You keep saying Chrome is a hog, but everyone that switched to Chrome did so because of the contrary : it's blazingly fast and light. What do you feel Chrome hogs exactly ?

Looking at Chrome's task manager (yes, Chrome provides its own) you can see the break down :

View attachment 346066

As you can see, very very little CPU is being used. Memory is also quite low on the browser itself, it's the tabs which contain heavy Web 2.0 code (Facebook/Gmail) that are more memory heavy since these tend to dynamically load and reload a ton of stuff in the background, but then again, total seems to be low for a modern browser (I've seen Safari use Gigabytes of memory).

Also notice how the Flash plugin is missing. It is enabled though. Chrome only loads it if a web page asks for it. So just by virtue of Flash being there, it's not using any ressources at all. Making it completely on-demand with FlashBlock prevents it from loading unless you want to on top. So you can't even use that to say "Chrome is a hog".

Again, the more you talk about this, the more I'm thinking there's either something very wrong with your computer, or you just don't understand modern OSes and how they work exactly. Hence I'm left puzzled by how you think you are tooled to even be making the claims you are making.

Flash in Chrome is more CPU-intensive than Flash on other browsers. It just uses more for some reason. Disabling Flash will sure use less CPU, but that is not an option if you want to view Flash. Oh, and it sometimes randomly pauses YouTube videos for some reason (on my computer, my brother's, my friend's, and others).

And you call that low RAM usage??? I don't think I've ever used that much RAM in Safari before. Actually, this could be from the OS throwing more RAM at it because more is available, which it does. I have 1GB of RAM, so I'm assuming you have more since this is the weakest MacBook. The real test is to open the same number of tabs on each browser on the same computer with no extensions with nothing else running. I tried that a few times, and Chrome always used more.