Tested on Mac OS 10.12.1 Sierra using chrome stable M54 #54.0.2840.87 and followed the steps:

1. Opened the chrome , loaded keePassX in the background .
2. In NTP, opened "cnn.com" and the page is loaded as expected.

Attaching screencast for reference.

@pricead-- could you please confirm whether the steps followed are correct?

Could anyone from MTV team  look into the this, as "AMD Radeon HD 6490M" GL is not available in In-house HYD team.

Thanks !

Deleted: 661596.mp4 682 KB

	661596.mp4 682 KB View Download

+erikchen

Unable to reproduce this on Mac OSX 10.12.1 - 54.0.2840.87 
The machine i tried it on has dual GPU Cards : AMD Radeon HD 6490 & Intel HD Graphics 3000 Open GL

Steps Followed:
1. Launched Chrome and navigated to youtube.com
2. Launched KeePassX application.
3. Changed the gfxcard to use AMD Radeon manually as it did not switch automatically on the machine i tested.
4. Navigated to www.cnn.com and it loaded fine.

Please let me know if i missed any steps.

Same problem here:

 54.0.2840.98 (64-bit) on OSX 10.11.6 El Capitan

MacBook Pro (17-inch, Early 2011)
Graphics: AMD Radeon HD 6750M 1024 MB

After using Chrome for a couple of hours two problems happen (at the same time):
- New tabs will open showing the same content of previous tab (while url is different).
- Black boxes appear on the screen for every screen update until whole screen is black

I tried the following:
- Disable all extensions
- Re-installing Chrome after downloading newest version from google website

This did NOT solve the problem

When the problem occurs I have to restart the browser. The problem is gone (for a couple of hours) until it happens again.

Deleted: Screen Shot 2016-10-28 at 2.14.44 PM.jpg 67.8 KB

	Screen Shot 2016-10-28 at 2.14.44 PM.jpg 67.8 KB View Download

Found an easy way to reproduce problem:

- Open a website in Chrome 
- Use an application which forces to use the fast GPU AMD HD 6750M (like adobe premiere)
- Then close Adobe Premiere -> GPU switches to discrete internal Intel graphics HD graphics 3000.
- Switch to Chrome
- Open a new tab in Chrome -> tab is showing black boxes.

Verified switch of GPU with gfxCardStatus app 

Problem also occurs when switching from discrete (internal Intel) to AMD 6750M.

- Open website in Chrome
(gfxCardStatus show Intel GPU is in use)
- Start Adobe Premiere
(gfxCardStatus show switch to AMD 6750M)
- Open a new tab in Chrome and navigate to website -> new tab stays blank

Deleted: Screen Shot 2016-11-24 at 1.01.54 PM.jpg 48.3 KB

	Screen Shot 2016-11-24 at 1.01.54 PM.jpg 48.3 KB View Download

Can you go to about:gpu, print the result to a PDF, and attach it?

We'll try to reproduce this on the same machine locally.

Here is the output of about:gpu (attached as PDF)

Deleted: chrome___gpu.pdf 110 KB

chrome___gpu.pdf 110 KB Download

From #11:
GPU0 VENDOR = 0x1002, DEVICE= 0x6741 *ACTIVE*
GPU1 VENDOR = 0x8086, DEVICE= 0x0126

 Issue 666887  has been merged into this issue.

I am experiencing these intermittent problems on my MacBook Pro Late 2011 15" with AMD Radeon HD 6770M 1024 MB
Intel HD Graphics 3000 512 MB
Chrome Version 54.0.2840.98 (64-bit)
One additional symptom I have is that when I use Cmd-Tab to switch applications, Chrome does not appear as one of the applications despite it being open. This seems to be occurring even when tabs are rendering properly, and I can workaround problem by closing Chrome and restarting it. Not sure if this is a different bug or additional symptom.

I've been experiencing this on my 2012 Macbook Pro and a very similar issue on my Windows 10 Pro desktop as well for at least the past 4-6 weeks now. The issues are not exactly the same but seem very similar so I will post info for both platforms here. I think they might be related.

What happens on the Macbook is described by previous reporters pretty well. My laptop has a NVIDIA GeForce GT 650M and Intel HD Graphics 4000. Full specs here:

https://support.apple.com/kb/sp694?locale=en_US

I'm running the latest MacOS Sierra.

On my desktop (Win 10 Pro), Chrome works for for a seemingly random period of time and then something happens that corrupts Chrome until I restart it. I haven't been able to connect anything I'm doing to this issue. Chrome simply appears to become corrupted under normal usage over a period of hours.

What happens is slightly different from the Mac: after a period of normal usage, when I open a new tab all that displays is a grayish blank page. If I try to enter a URL I get a file folder with a frown face in the tab itself (image attached with the other tabs blanked out, since they contained private info).

When I switch to a tab I already had open and try to enter a new URL there, nothing happens. Chrome doesn't appear to load the new page at all. I've also tried opening a new window in incognito mode - same deal. None of the settings pages load either. Nothing works to correct this issue except restarting Chrome.

So far, I've tried:

- Totally reinstalling Chrome
- Wiping my profile and re-syncing it
- Running the Chrome cleanup tool
- Disabling hardware acceleration
- Disabling all extensions

I'm not running any third party AV on the desktop (just Windows Defender) and nothing on the Mac.

Here is a post from a guy who started seeing this happen in his office:

https://productforums.google.com/forum/?utm_medium=email&utm_source=footer#!msg/chrome/Tf4BehUkWL8/lnaCJ-oEBQAJ

He says disabling Remote Desktop Services and rebooting fixed the issue for his users. I haven't tried this yet myself but will likely do this now as I use Chrome extensively for work and this has been becoming an increasingly painful issue for me. I'm hoping this gains traction and gets fixed ASAP.

I'll also attach an about:gpu for my desktop. The interesting thing is that it also has a GeForce card (GTX 670) which is also a 2012 graphics card. Perhaps there is an issue with this particular GeForce platform.

Let me know if you need any additional info.

Deleted: ChromeBlackScreen.png 10.7 KB

	ChromeBlackScreen.png 10.7 KB View Download

Deleted: chrome___gpu.pdf 142 KB

chrome___gpu.pdf 142 KB Download

I am experiencing this issue too, on a MacBook Pro Late 2011 and am running macOS Sierra and the latest version of Google Chrome (currently 55). As requested, I have attached the output of chrome://gpu/

Deleted: gpu.pdf 79.3 KB

gpu.pdf 79.3 KB Download

I haven't seen this happen on MacOS or Windows for several days now.

Currently running Chrome version 55.0.2883.87.

Still experiencing the issue in 55.0.2883.87 (64-bit).

Still experiencing this on 55.0.2883.87 and the latest MacOS 10.12.2

This is a really jarring issue. Can also confirm that this is still happening on 55.0.2883.95 and macOS 10.12.2.

What I found to be helpful is to kill the "GPU process" with the Chrome task manager after the system switches back to the integrated graphics. The tabs just reappear after that.

Deleted: gpu.html 754 KB

gpu.html 754 KB View Download

Link to additional reports of this issue: https://productforums.google.com/forum/?utm_medium=email&utm_source=footer#!msg/chrome/pwoOKvS7-cg/2dC_wYk_DgAJ

Adding Needs-TestConfirmation.

Can we try to find a machine among our manual test machines to reproduce this?

Spoke too soon. Chrome on Windows is still becoming corrupted and showing a blank grey screen with a frown face in the tab (until restarted). I haven't seen this issue on my Macbook Pro for a while but I don't use that all that often either. This is definitely a very jarring issue. 

I have a user on a macbook air running 55.0.2883.95 on OS 10.12.2 having the same issues I am seeing reported here. With the exception that when chrome is running on his monitor connected via Thunderbolt display he has no problems it is when he moves it to one of the screens connected via USB to DVI that it happens. I can watch it happen trough a go to assist session so I do not belive it is the fault of the cables. if he leaves a screen between the two monitors half of the browser window is white the other acts as normal. We have unchecked the box for graphics acceleration to see if that would help as well but it did not. 

I noticed I have the same problem with Opera, besides Chrome:

1. opening Opera and some sites
2. connecting external screen
3. opening new tab
4. the tab is either completely black or white
5. If I quit and open again then it works fine, until I disconnect external monitor.

Version:	42.0.2393.517 - Opera is up to date
Update stream:	Stable
System:	Mac OS X 10.12.2 64-bit

MacBook Pro (15-inch, Early 2011)

This happens _every_ single time my late 2011 17" MBP switches between discrete and integrated graphics. I have to restart chrome twice a day, once when I plug into my thunderbolt display at work in the morning (switch to discrete), and once when I unplug from said screen in the evening (back to integrated).

This is extremely frustrating but has just become part of my daily routine.

I'm on El Cap and I believe this started happening around Chrome v54 (I'm on dev channel v57 right now).

See this product support thread: https://productforums.google.com/forum/#!topic/chrome/pwoOKvS7-cg;context-place=topicsearchin/chrome/category$3Amac

There seems to be something about 2011 vintage MBPs that exacerbates this. Radeon perhaps?

Attached is the output from my chrome://gpu

Deleted: gpu.html 133 KB

gpu.html 133 KB View Download

I get it a lot on my late 2011 15" MBP, Sierra 12.10 (16A323)
Chrome Version 56.0.2924.87 (64-bit)

Happens predictable when I open GIMP.

I reached out to Test to ping again on finding machines where this reproduces. We really need to get on this as it's a P1. 

I couldn't find the exact GPU machine (AMD Radeon HD 6750M 1024 MB), however i was trying to reproduce this on one of the GPU (AMD Radeon HD 6490M 1024 MB) for OS X# 10.12.3 MacBook Pro (15-inch, Early 2011) and couldn't succeed. The pages were rendering w/o any issues even i was switching b/n two Graphic cards (AMD Radeon HD 6490M 1024 MB & Intel HD Graphics 3000 OpenGL Engine) as per c#8.

chrome://gpu:
==============
GPU0	VENDOR = 0x1002, DEVICE= 0x6760
GPU1	VENDOR = 0x8086, DEVICE= 0x0116 *ACTIVE*

Could someone on this thread please confirm, any of you still seeing this issue with OS X 10.12.3 as well? Just curious to know whether 10.12.3 has made some fix or not.

Thank you!

I have a Early-2011 MacBook Pro with an AMD Radeon HD 6490M (256mb
version), and I'm still experiencing this issue on 10.12.3 and Chrome
56.0.2924.87 (64-bit).

I still have the issue on 10.12.3
Late 2011 MBP, with MAD Radeon HD 6770M, 1024Mb (includes Intel HD Graphics 3000 in low power mode)
Chrome Version 56.0.2924.87 (64-bit)

I have MacBook Pro (17-inch, Early 2011) with Graphics: AMD Radeon HD 6750M 1024 MB with OSX 10.11.6 and Chrome 56.0.2924.87 (64-bit) -> Still a problem.

I do not think the problem is in OSX, because Firefox (v 50.1.0) and Safari (10.0.3 (11602.4.8.0.1)) do not have this problem. The problem only occurs with Chrome.

Still happening with macOS 10.12.3, Chrome  56.0.2924.87 (64-bit) on Late 2011 MBP with Intel HD Graphics 3000 512 MB and AMD Radeon HD 6770M 1GB

> I reached out to Test to ping again on finding machines where this reproduces. We really need to get on this as it's a P1. 

Should we acquire one of these machines since test does not have one?

Thank you for all the updates.  Finally i could able to repro this on "AMD Radeon HD 6490M 1024 MB" on OS X Sierra.  I will try to see if i can provide the narrow bisect.

Great, thanks for the repro, Mano!

Assigning to Mano for the bisect.

Can confirm that this issue is NOT limited to Mac OS Sierra, but also El Capitan. Here are my system specs:

Model Name:	MacBook Pro

  Model Identifier:	MacBookPro8,3 (late 2011, 17-inch)
  Processor Name:	Intel Core i7
  Processor Speed:	2.5 GHz
  Number of Processors:	1
  Total Number of Cores:	4
  L2 Cache (per Core):	256 KB
  L3 Cache:	8 MB
  Memory:	16 GB
  Boot ROM Version:	MBP81.0047.B2D
  SMC Version (system):	1.70f5
  Hardware UUID:	B1A6ACA1-449C-599A-9F04-1AEDDE66B229


System Software Overview:

  System Version:	OS X 10.11.6 (15G31)
  Kernel Version:	Darwin 15.6.0
  Boot Volume:	ELC-SYS
  Boot Mode:	Normal
  Secure Virtual Memory:	Enabled
  System Integrity Protection:	Enabled
  Time since boot:	2:19


Graphics/Displays:

Intel HD Graphics 3000:

  Chipset Model:	Intel HD Graphics 3000
  Type:	GPU
  Bus:	Built-In
  VRAM (Dynamic, Max):	512 MB
  Vendor:	Intel (0x8086)
  Device ID:	0x0126
  Revision ID:	0x0009
  gMux Version:	1.9.24

AMD Radeon HD 6770M:

  Chipset Model:	AMD Radeon HD 6770M
  Type:	GPU
  Bus:	PCIe
  PCIe Lane Width:	x8
  VRAM (Total):	1024 MB
  Vendor:	ATI (0x1002)
  Device ID:	0x6740
  Revision ID:	0x0000
  ROM Revision:	113-C0170L-573
  gMux Version:	1.9.24
  EFI Driver Version:	01.00.573
  
Displays:
  Color LCD:
  Display Type:	LCD
  Resolution:	1920 x 1200
  Pixel Depth:	32-Bit Color (ARGB8888)
  Main Display:	Yes
  Mirror:	Off
  Online:	Yes
  Built-In:	Yes


From all the literature out there it appears to have been introduced at some point by a Chrome update. Seems like a Chrome fix is the only true fix, as disabling HW Acceleration is like asking us to choose between our legs or our eyes :(

I'm able to reproduce this issue on Latest Canary#58.0.3008.0, Dev#58.0.3004.3, Beta#57.0.2987.37 and Stable#56.0.2924.87 versions of Chrome for Mac OS X 10.12.3.

Good#53.0.2744.0
Bad#54.0.2786.0

Here is the narrow bisect.
https://chromium.googlesource.com/chromium/src/+log/e973b43fd81bb5066ec3c95034e8290cde5068eb..2b6875bfd954d336432702512d1ffe4c83343d2b

piman@, Can you please take a look if something is related to https://chromium.googlesource.com/chromium/src/+/89d5065c918e90c43c98c75e3e3868f929d7af93 ?

Thank you!

piman@ hasn't been to crbug in over a month. Should we find a different owner?

ericrk@/ccameron@, Could one of you please take a look and find a owner if possible?

Thank you!

My guess is that this started with https://chromium.googlesource.com/chromium/src/+/0b113ce33d9d553b105fa68a2b2b554a93de43e9

which blacklists HD3000. All the machines reporting this issue have HD3000s, and I'd suspect that there is an issue when transitioning from the blacklisted GPU to a supported discrete GPU. I'm going to update the blacklist to handle this better (blacklist regardless of whether the HD3000 is primary or secondary).

Can you elaborate on your comment in c#44? Are you updating the actual blacklist or the machinery that enforces the blacklist settings? I'm interested in the exact change you are making here as well as what we are doing/can do to avoid this kind of problem in the future.

You are totally right ericrk@ ! I used the flag --ignore-gpu-blacklist and this fixed the problem in Opera. I can now switch between integrated and high perf. graphics without causing the browser those visual glitches. 

I haven't tried in Chrome, but the symptoms were the same for Opera and this fixed it.

I can confirm that setting the flag #ignore-gpu-blacklist (Override software rendering list) fixes the problem in Chrome 56.0.2924.87 (64-bit) in OSX 10.11.6 on a MacBook Pro (17-inch, Early 2011. When the flag is set I can switch between integrated and discrete videocard (using the tool gfxCardStatus) and Chrome is still able to load new tabs.

re comment #45 - good point on preventing this going forward - I'll both update the blacklist, as well as adding a DCHECK to the logic to ensure that  blacklist entries on Mac always apply to both GPUs (I think this is a mac-specific issue, we haven't seen similar issues on Windows - I'll double check this though).

Is it possible for end user to set this flag? Is there a recipe for doing
it? Thx!

Joe Murray, PhD
President, JMA Consulting
joe.murray@jmaconsulting.biz
skype JosephPMurray twitter JoeMurray
416.466.1281

Joe, you can set the flag manually using the following in url: chrome://flags
and then enable the following option: #ignore-gpu-blacklist (Override software rendering list)

Is it possible that this issue is related to this glitch I'm still seeing on Windows?

https://productforums.google.com/forum/?utm_medium=email&utm_source=footer#!msg/chrome/Tf4BehUkWL8/lnaCJ-oEBQAJ

The only reason I ask is because as far as I can recall, I started seeing this both on my 2012 Macbook Pro AND my Windows desktop with GTX 670 at the same time (Oct/Nov of 2016). The GTX 670 was purchased in 2012 as well. I posted details above.

brendon@, I assume your Windows desktop is a single-GPU setup? If that's the case, this is very likely a different issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d0f1bb54f393f66586c7e85ccddfb314a146be61

commit d0f1bb54f393f66586c7e85ccddfb314a146be61
Author: ericrk <ericrk@chromium.org>
Date: Wed Feb 15 08:01:40 2017

Ensure that HD3000 blacklist on Mac covers "any" multi_gpu_category

HD3000 was previously blacklisted for all GPU features. However, this
only applies when that card is the primary GPU. We've seen issues with
this in the past - moving this to the "any" category, ensuring that we
keep consistent GPU blacklist status around GPU switches.

BUG= 661596 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2691013003
Cr-Commit-Position: refs/heads/master@{#450634}

[modify] https://crrev.com/d0f1bb54f393f66586c7e85ccddfb314a146be61/gpu/config/software_rendering_list_json.cc

I would also question blacklisting the HD3000. I never had problems with it on my mac.

System:	Mac OS X 10.12.2 64-bit

MacBook Pro (15-inch, Early 2011)
Intel HD Graphics 3000 512 MB + AMD Radeon HD 6750M (1024MB)

Also, remember the user feedback you got when doing https://codereview.chromium.org/2076443002 (see the comments)

Blacklisting both GPUs on macs with HD 3000 will make users wonder why hardware acceleration doesn't work on Chrome. Lucky ones like me who followed this thread will set the flag --ignore-gpu-blacklist but what about all the others?

Again, never had any problems in Chrome with HD 3000 or my current setup before you blacklisted the GPU. 

As probably expected: now I have enabled the --ignore-gpu-blacklist flag chrome is crashing WebGL. In Google maps i get the following error: Rats. WebGL hits a snag.

Jeroen, 
Me too. Late 2011 macbook pro 15"

I agree that it's not ideal to blacklist both cards on a multi-gpu setup, when we really only need to blacklist one. I've open a follow-up bug to try to investigate why the switching isn't working - crbug.com/692632.

Regarding why we blacklisted HD3000, we found that that card was responsible for a large portion of our GPU crashes on Mac. It's a case where we're trying to strike a balance between performance and stability. I'll look into this a bit more as well.

Same issue-
I wish I had found this page before I went ahead and did a completely clean install of Sierra because of the issue!

MacBook Pro (17-inch, Early 2011)
2,3 GHz Intel Core i7
8 GB 1333 MHz DDR3
AMD Radeon HD 6750M 1024 MB
Intel HD Graphics 3000 512 MB

So, is it safe to do this:

"Joe, you can set the flag manually using the following in url: chrome://flags
and then enable the following option: #ignore-gpu-blacklist (Override software rendering list)"

Or should I wait.
Thanks

Re #61, the safest option is to run Chrome with the --disable-gpu flag. If you want to try using --ignore-gpu-blacklist, that may work for you, but may also decrease Chrome's stability.

Same issue, also on Sierra.  MacBook Pro (15-inch, Late 2011).  Looking forward to a fix!

Your change meets the bar and is auto-approved for M57. Please go ahead and merge the CL to branch 2987 manually. Please contact milestone owner if you have questions.
Owners: amineer@(clank), cmasso@(bling), ketakid@(cros), govind@(desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Please merge your change to M57 branch 2987 before 5:00 PM PT Friday (02/17), so we can pick it up for next week Beta release. Thank you.

The above fix (per c#54) has been verified on Latest Canary#58.0.3014.0 for Mac OS X 10.12.3 and chrome is rendering with out any issues.

Thank you!

Not sure why this wasn't auto updated. This was merged in crrev.com/2704563004

I can confirm that this is fixed for me in the latest Canary (Version 58.0.3015.0 canary). I look forward to the beta version next week. Thank you everyone!

Please merge your change to M57 branch 2987 before 5:00 PM PT Monday (02/20), so we can pick it up for next week Beta release. Thank you.

This was merged into 57 in crrev.com/2704563004

Applying "merge-merged-2987" label for #71.

The above fix has been verified on beta #57.0.2987.74 for Mac OS X 10.12.3 (HD Graphics 4000) and chrome is rendering without any issues.

As per comment#54, this needs a HD 3000, Can MTV team take a look into it on this configuration.

Thank you!

Had an offline chat with ericrk@ and we need a Mac machine which has HD 3000 and a discrete GPU as well(Switchable) in order to verify the issue, I have looked up the GPU inventory and can just find a machine with HD 3000 

pricead@ can you please check and let us know if the issue got fixed in latest Chrome Beta(A version higher then 57.0.2987.54)

@pbomm I just received an update to 57.0.2987.74 beta and everything seems to be working fine on this version. No signs of the bug after forcing a switch between integrated and discrete (and vice versa) and opening new pages in Chrome. Thanks again!

Thank you so much  pricead@ for verifying.

 Issue 656675  has been merged into this issue.

 Issue 692189  has been merged into this issue.

Has this bug creeped back in? This went away for a long time and now is back. Early 2011 Macbook Pro, macOS Sierra 10.12.6 and latest Chrome 62.0.3202.75.

Hi Mano,
Can you please check if this has regressed?
Thanks!

I'm not seeing this issue on Latest Chrome Stable#62.0.3202.75 for MBP (15-inch, Early 2011), 'Intel HD Graphics 3000' & 'AMD Radeon HD 6490M'. Just wondering, could someone else also seeing this?

Thank you!

To be clear, it's never been consistent for me - just a once in a while thing. Which I know makes it tough.

I agree with Comment 81 by futzlar, I am now seeing this issue following recent updates to High Sierra and Chrome 62, which occurred at roughly the same time. Now several times a day I will experience pages failing to render in existing or new tabs, although I can watch the background activity when I profile the page load in developer options. Like futzlar, I cannot always make this happen, but it will undoubtedly happen when I don't want it to. Restarting Chrome is effective for a short while.

Chrome Version 62.0.3202.89, Mac OS 10.13.1, MacBook Pro 17, Early 2011 (AMD Radeon HD 6750M 1GB, Intel HD Graphics 3000 512 MB)

futzlarson@ and marco@, if possible, are you able to paste the contents of about:gpu next time you see this issue (ideally right after). Navigate to about:gpu and save-as "website, complete".

Actually haven't seen it in a number of days now but yes, can do.

OK, I tried the last couple days to make it happen, so of course it didn't...

..until I went ahead and applied the update to 62.0.3202.94 tonight. I had left Google Earth Pro running, so my AMD GPU was running. When Chrome restarted, I looked at the about:gpu and all the hardware acceleration was running except rasterization. Everything seemed to be working fine. Then I quit Earth to try to kick it back down to the intel GPU, but Chrome was keeping the AMD engaged. So I quit Chrome and OSX dropped back to the intel GPU. Then restarted Chrome, looked at about:GPU, saw all the red from disabled hardware acceleration, maps.google.com was 2D only, as expected. Then started google earth, AMD GPU fired up, went back to about:GPU and saw it (almost) all green. Then tried maps.google.com, which tried 3D but didn't work (just saw a google logo background), and on opening new tabs the viewport didn't render or just kept displaying the previous tab. Ran a fresh about:gpu and saved it (although I couldn't see it myself), and have attached that now.

I see that .94 had some updates to rendering, perhaps that changed?

Deleted: gpu.htm 51.9 KB

gpu.htm 51.9 KB View Download

Actually I'm just trying maps.google.com again in 3D using the discrete AMD and it's working great now. I'd certainly prefer the hardware acceleration in Chrome, as long as it's relatively stable.

macro@: Chrome had intended to disable all GPU functionality with the Intel HD 3000 GPU (which is the integrated GPU on your system) because of serious instability issues in the driver. That's the reason WebGL works on your system when the AMD GPU is in use.

Same issue here again. It was gone until I updated to OSX High Sierra last week, could be a coincidence as some report that it started with Chrome 62.0.*

Same symptoms as before, pages stop rendering, show content from other tabs, white or black screens but in the background the page got loaded.

Chrome got almost unusable now, it feels much worse then the last time the bug occurred :/

Apparently the symptom happens when the GPU switches from the ATI back to the Intel one.
Attached are two about:gpu pages, one right when I start chrome and the other one after it switches from ATI to Intel and the bug occurs.

Chrome/62.0.3202.94
MacBook Pro (15-inch, Early 2011)
2,2 GHz Intel Core i7
AMD Radeon HD 6750M 1 GB
Intel HD Graphics 3000 512 MB

Deleted: gpu_chromestarted.html 46.5 KB

gpu_chromestarted.html 46.5 KB View Download

Deleted: gpu_switchfromATItoINTEL.html 201 KB

gpu_switchfromATItoINTEL.html 201 KB View Download

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/0c5fd85b836bb7216a5acc1ed9e463ff26c436ae

commit 0c5fd85b836bb7216a5acc1ed9e463ff26c436ae
Author: Kenneth Russell <kbr@chromium.org>
Date: Tue Nov 21 01:31:29 2017

Restore multi_gpu_category=any for Intel HD 3000 blacklisting.

This was apparently lost during the switch to use JSON files for the
blacklists.

BUG= 661596 

Cq-Include-Trybots: master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: I3c4658d2c8d14f2f6e6b8ce9a591693193780714
Reviewed-on: https://chromium-review.googlesource.com/778187
Reviewed-by: Eric Karl <ericrk@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#518031}
[modify] https://crrev.com/0c5fd85b836bb7216a5acc1ed9e463ff26c436ae/gpu/config/software_rendering_list.json

Commenters: this bug did indeed regress during a recent large rewrite of how graphics driver bugs are tracked. It'll be worked around in the next Chrome Canary.

Note that the workaround will unfortunately disable all GPU acceleration on your machines. This is how it worked previously.

Unable to reproduce the issue on Mac Os 10.12.6 & 10.13 high sierra with discrete GPU only using chrome stable M62 #62.0.3202.94 and followed steps mentioned in comment #3.

Attached screencast for reference.

@MTVTeam-- Could someone from MTV team look into this as in-house team does't have the Intel HD 3000 GPU .

Thanks!

Deleted: 661596 (1).mp4 5.6 MB

	661596 (1).mp4 5.6 MB View Download

Is this bug repro on M63? If yes, do we need a merge for cl listed at #92 to M63?

Yes, this bug will reproduce in M63 as we seem to have regressed the blacklist entry in M62. Shall we merge the change of the blacklist entry back to M63?

This bug requires manual review: We are only 13 days from stable.
Please contact the milestone owner if you have questions.
Owners: cmasso@(Android), cmasso@(iOS), gkihumba@(ChromeOS), govind@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

To clarify: the change is safe and we're going back to the behavior in M58, M59, M60 and M61. M62 regressed.

Approving merge for listed at #92 to M63 branch 3239 based on comment #98 and per offline chat with kbr@. 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d41aea888d574eb8338a69a6d2648d9c145bdb00

commit d41aea888d574eb8338a69a6d2648d9c145bdb00
Author: Kenneth Russell <kbr@chromium.org>
Date: Tue Nov 21 19:56:01 2017

Restore multi_gpu_category=any for Intel HD 3000 blacklisting.

This was apparently lost during the switch to use JSON files for the
blacklists.

BUG= 661596 
TBR=kbr@chromium.org

(cherry picked from commit 0c5fd85b836bb7216a5acc1ed9e463ff26c436ae)

Cq-Include-Trybots: master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: I3c4658d2c8d14f2f6e6b8ce9a591693193780714
Reviewed-on: https://chromium-review.googlesource.com/778187
Reviewed-by: Eric Karl <ericrk@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#518031}
Reviewed-on: https://chromium-review.googlesource.com/782648
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/branch-heads/3239@{#555}
Cr-Branched-From: adb61db19020ed8ecee5e91b1a0ea4c924ae2988-refs/heads/master@{#508578}
[modify] https://crrev.com/d41aea888d574eb8338a69a6d2648d9c145bdb00/gpu/config/software_rendering_list.json