Display Issue - Horizontal blue lines in application windows

Unfortunately, the above interim solution does not work consistently. See follow-up comment posted there (#189).

Apple, Safari is now affected too. This is not a one-off problem with just my hardware – the glitches are being reported across multiple users on the mbp 16".

To be clear, resetting NVRAM doesn't work – this should not have been accepted as the 'answer' for this question.

Restarting resolves the issue, only for it to come back again.

I'll share links here again for anyone wanting to understand the extent of this issue:

https://support.google.com/chrome/thread/39447594?hl=en

https://bugs.chromium.org/p/chromium/issues/detail?id=1068170

This bug has made it near impossible to use my Macbook for work. I'm a software engineer that uses vs code (electron), slack (electron), and all flavors of browsers (mostly chromium based).

I find myself having to restart constantly in hopes that I don't have green/blue lines everywhere...

Even the chromium "fix" of spam restarting the gpu process for those apps seems to be hit or miss. Sometimes on the 3rd or 4th or 7th time the artifacts will disappear. Regardless, after spam restarting those processes, the app never seems to work right until closing and opening again which, you guessed it, caused the artifacts to reappear. So, I'm back to restarting the machine.

From the Chromium bug tracker (which is now locked):

https://bugs.chromium.org/p/chromium/issues/detail?id=1068170#c195

Everyone: Apple has indicated that they have created a more reliable set of steps to reproduce this in-house and continue to investigate it.

Whose court is the ball in now? I do not know.

Perhaps it's just us software devs complaining about this issue. Seems that way. I believe software devs encounter this more often as we are coding, testing and leaving our WebGL views open and running on our browser tabs all the time. The problem is particularly easy to reproduce if you have intensive WebGL operations running (say three.js/babylon.js/pixi.js), let your machine go to sleep and wake it up after an hour (or overnight).

Interestingly, the Google Chrome support page is now rife with end-users claiming the problem was 'worked around' when they disabled H/W acceleration – as I've mentioned multiple times above, not a permanent solution given the drastic reduction in WebGL graphics rendering performance.

Some are even claiming success in resetting NVRAM – or what I call the slap-the-TV-in-case-of-poor-reception fix. No, it doesn't work. Would have been great if it did! I'm wondering why folks would post things like that and call it a day. It is more likely that resetting NVRAM simply forces you to restart the computer, but the problem will eventually come back again.

Anyway, I'm glad I'm not the only one facing a similar situation that is impacting how we work. The Chromium GPU process restarts no longer work reliably for me – I just force a reboot. Incidentally all similar Chromium apps have the "GPU Helper" process that you can also attempt to spam kill but it hasn't worked for me.

I just wish more will step forward to echo these similar issues. Thanks for chiming in.

chuank wrote:

From the Chromium bug tracker (which is now locked):

https://bugs.chromium.org/p/chromium/issues/detail?id=1068170#c195

Everyone: Apple has indicated that they have created a more reliable set of steps to reproduce this in-house and continue to investigate it.

Whose court is the ball in now? I do not know.

Perhaps it's just us software devs complaining about this issue. Seems that way. I believe software devs encounter this more often as we are coding, testing and leaving our WebGL views open and running on our browser tabs all the time. The problem is particularly easy to reproduce if you have intensive WebGL operations running (say three.js/babylon.js/pixi.js), let your machine go to sleep and wake it up after an hour (or overnight).

Interestingly, the Google Chrome support page is now rife with end-users claiming the problem was 'worked around' when they disabled H/W acceleration – as I've mentioned multiple times above, not a permanent solution given the drastic reduction in WebGL graphics rendering performance.

Some are even claiming success in resetting NVRAM – or what I call the slap-the-TV-in-case-of-poor-reception fix. No, it doesn't work. Would have been great if it did! I'm wondering why folks would post things like that and call it a day. It is more likely that resetting NVRAM simply forces you to restart the computer, but the problem will eventually come back again.

Anyway, I'm glad I'm not the only one facing a similar situation that is impacting how we work. The Chromium GPU process restarts no longer work reliably for me – I just force a reboot. Incidentally all similar Chromium apps have the "GPU Helper" process that you can also attempt to spam kill but it hasn't worked for me.

I just wish more will step forward to echo these similar issues. Thanks for chiming in.

+1

I guess there isn't a definitive solution yet? Is Apple waiting for people to just give up on this?

Disabling Hardware Acceleration is NOT a solution, it's like having an Hybrid car and turning off the engine just to run on electric motor.

I am fed up with this situation and it is unacceptable. I can't continue working like this.

For what it's worth, I have not had the issue occur for a while. I have a 3 monitor set up using a CalDigit docking station on my Macbook Pro 16in. 1 monitor is a 4k monitor, the other is a standard monitor and the 3rd is the macbook display. I am not positive if the events are related but when I am done for the day I have been keeping my macbook connected to the docking station and started closing my macbook to make it go to sleep. It's probably been a good month or so since I've had the blue line issue. I thought I would share. Thanks

Totally the opposite for me, I used to have it less before but now it's happening every time my laptop wakes up from sleep. I also have CalDigit TS3 docking with 2 external monitors. It happens regardless laptop is connected to docking station or not.

To concur with @fexnok – the issue is present regardless of the presence of external displays. The glitches appear on all displays when it triggers.

My projects involving WebGL have concluded, which corresponds directly to the glitches not occurring as often, but as I have mentioned before multiple times above, all you have to do is to keep a WebGL-intensive website running (an easy test is to visit https://www.shadertoy.com/), let your mac go to sleep and upon waking up, there's a very high chance of seeing the glitch.

There is also the chance that software publishers are aware of the glitch and have simply elected not to use WebGL views in their applications to prevent this in their products. However, Chromium-based desktop apps have a dead giveaway – just toggle the Developer Tools window (if it's available) and you'll see the glitches when it triggers.

As it stands today, this is still a critical issue. To date, since April 2020, we still do not know if this is hardware or software-related. It's not on Chromium either, other applications have exhibited similar artifacts (Safari).

This isn't about putting the blame – it's about identifying and fixing the problem so that we can get on with our work effectively.

PS it's really unfortunate that this thread has an 'answer' which isn't true, and throwing off folks who are encountering the same problem. It will be great if Apple can unflag resetting NVRAM as the solution. Also, the way the thread shows just that first page with a purported 'solution' is misleading. We've hit 10 pages of comments – unfortunately most of them being "disable H/W acceleration fixed it for me!"-type responses. Again, not helpful or valid for those of us tearing our hair out on this glitch.

Just noticed in today's glitch report. It was the usual 'trigger' – let the MBP go to sleep overnight, wake it in the morning to resume work, blue/green glitches.

This time though, it's not intensive WebGL views that manifested the glitch, but just standard UI webviews drawn via JS from a browser window. In other words, a website / web applications that likely utilised some form of graphics acceleration in JS.

Nothing new, just confirming and renewing the issue here – that this glitch is not an edge case limited to high-intensity WebGL graphics (they simply are more likely to trigger).

Hi @haffystyle, not that I know of.

A developer from Figma has posted more technical details to the Chromium bug tracker – linked to SVG rendering in the case of Figma. Glitches are also present in Safari although in my case the artifacts do not go away until a system restart is done:

https://bugs.chromium.org/p/chromium/issues/detail?id=1068170#c202

It has been almost ten months since the original issue was posted. This matter is still unresolved.

Owning the flagship 16'' laptop maxed out for over a year and dealing with this problem from day 1. All I have to say is that Apple drives me away slowly. I have not upgrade to Big Sur as the problem exists there too (+more). I don't care if the problem lies with Apple or Google Chrome or graphic card driver, all I care about is to be able to do my work without any OS headaches or I can always go back to Linux or Windows.

Resetting NVRAM/PRAM obviously does not work. Only workaround -and not a fix in my case- is to disable hardware acceleration to all apps builded with Electron where that is doable (eg Visual Studio Code) and Chrome. For those of us who need to use hardware acceleration we are doomed. Thank you Apple for ignoring us one way or another!

Yup hardware acceleration is the issue. Had this on Firefox with the lucid chart plugin of confluence. Deselected performance settings and hw acceleration and it is fixed. Issue began with a Catalina security update.

So I’ve also had this issue quite frequently lately. I noticed today that it only seems to happen when the battery is drained. If I am watching Twitch on 1 monitor and working on another the monitor pass through charging isn’t enough juice apparently. If I’m not paying attention the battery can drop to like 5-10%. The kernal_task will skyrocket, computer slows down, and then this issue happens. Not sure if this is the only cause but figured I would post it. I plugged my actual charger in today and didn’t have the issue again. Will update if it doesn’t seem to resolve moving forward.

I think it's critical not to describe this issue as being 'fixed' when the proposed solutions are just workarounds. This is not to disregard the helpful advice others in the community have been providing though. However, the problem is still there, and there's an elephant in the room.

Disabling HW acceleration sidesteps the issue. The combination of software/hardware issues triggering the graphics corruption is still there, you are just sticking duct tape over the leaky hole. For those of us who rely on the machine to write WebGL-intensive code and projects, this isn't a satisfactory solution, and I'm glad I'm not the only one saying this here.

Automatic Graphics Switching is similar too – I would venture that toggling it off/on again gives the graphics subsystem a way to re-initialise certain portions of the hardware acceleration without rebooting. It does not work for me – I wish it could.

At the very least it seems like Apple folks are getting cc'ed into the loop at the Chromium bug tracker

I tried to find Metal in the flags but it wasn't there.

I did, however, discover that if I went to (three dots)/More Tools/Task Manager in Chrome, and then sorted the tasks there by name, force-quit Renderer, Spare renderer, and GPU Process, the page refreshes, the corruption cleared up and stayed cleared. It's been about 30mins and it's still clear, but I'll report back if it fails again.