We should all be grateful that they create an endless supply of products on the Apple Refurbished store.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iMac 2020, 5700XT - graphical glitch?
- Thread starter torifile
- Start date
- Sort by reaction score
We should all be grateful that they create an endless supply of products on the Apple Refurbished store.
Question, wouldn’t offering evidence of this glitch to apple warrant a return under the normal 1-year warranty? Or is it restricted to replacements/repairs after the 14-day return window?
It would appear this is a combination of hardware and potentially software as well, but mainly I think this can be attributed to the architecture of the cards using DSC (display stream compression). The reason you do not see this on any previous gen cards (Vega series) is because they do not attempt to compress the display signal. These Navi’s cards are flaky on Windows too. It’s not just specific to the Mac. Just an FYI why you may not see this on certain resolutions as it alters the amount of display stream compression.
I would assume with the length of time Apple has had these GPUs they would’ve thoroughly tested them and are aware of the issues / potential drawbacks.
I would assume with the length of time Apple has had these GPUs they would’ve thoroughly tested them and are aware of the issues / potential drawbacks.
The connection to the iMac display is not using DSC. It's dual DisplayPort 1.2 which doesn't even support DSC.
Now, if you had a DisplayPort 1.4 MST hub and connect a DisplayPort 1.2 display, then the GPU may output DSC to the MST hub, which then decompresses it to send to the DisplayPort 1.2 display. If you were running Windows, DSC will let you use more displays whether they support DSC or not since the MST hub will do the decompression if necessary.
Just received a new system I configured with the Vega II for comparison in my MacPro.
W5700X driving dual XDR displays = completely glitched out on 2 systems I replaced
Vega II driving dual XDR displays = 100% perfect, no corner window glitching, white lines, or blocky corruption as seen with the W5700X in setup screens.
I've documented the issue across two W5700X machines now. Got them. Apple can't deny the issue any longer
W5700X driving dual XDR displays = completely glitched out on 2 systems I replaced
Vega II driving dual XDR displays = 100% perfect, no corner window glitching, white lines, or blocky corruption as seen with the W5700X in setup screens.
I've documented the issue across two W5700X machines now. Got them. Apple can't deny the issue any longer
Attachments
Are you able to verify the iMac isn't using DSC to drive the display? Something is up with these Navi cards regardless. The DSC hardware and software setup would be my best guess.
That was my gut instinct two or three weeks ago too. I don’t know how the iMac display is being driven (I don’t think any of the teardowns looked at that and I don’t think we’ve seen a teardown of a 5700XT model) and it very likely could still be using DP 1.2 — but I have a feeling — based on gut instinct and nothing more — that DP 1.4/DSC could be part of this. That would also align with the lack of issues using an external display (though obviously would not explain the issues Mac Pro owners see with the W5700X and the XDR).
I guess the ultimate “what” doesn’t matter as long as it can be fixed. But the issues across this line of chips across Macs definitely indicates a driver issue or kernel issue of some sort.
The output of the AGDCDiagnose command will tell you how a display is connected to the AMD GPU. For iMac 5K, it will be dual DisplayPort 1.2 HBR2 without DSC.
For XDR, it will be single DisplayPort 1.4 HBR2 with DSC or dual HBR3 without DSC.
The lack of issues on an external display would actually support my theory, as a lot of external displays don't support DSC. The XDR display does indeed support and run using DSC on the 5500 and 5700 available on the Mac Pro.
@joevt while that may be true what your saying, there's still hardware involved in allowing DSC to be supported on all of these graphics chips. In order to support 2x XDRs on the iMac and MacBooks equipped with these Navi based GPU's, the DSC hardware is present regardless of whether the current display mode supports it.
Why else would basic 2D graphics output be difficult to figure out and have it running without glitches? A company like AMD has been making solid and visually accurate display outputs for decades - that's why to me, the major difference in this generation of cards we have with Navi is really only down to DSC support at that hardware and software level.
There is no way Apple could have known that the 5700XT would produce these graphic glitches. I know for a fact that Apple only tested the iMac all the non-Default resolutions. Also, rumor has it Apple already has a software fix for this. After you update to 10.15.7 next week, you'll notice that the option to select the Default resolution is gone. They deleted it. Now you are just forced to use one of the others. Problem solved. ?
Last edited:
Just because you declare everyone has the issue doesn't make it so. I stare at my screen for 9+ hours every day using Xcode, Photoshop, Word, Excel, Outlook, Safari, After Effects, Chrome, Terminal, you name it. I would notice glitching. Maybe there's some factor on my machine that makes it different, I don't know. But, contrary to your laughing at people, you are not correct because there are exceptions. I use a second monitor as well. Who knows, maybe that kills the bug in my particular set up.
External monitors are known to cure the problem (at least in most cases). Try without and report back? It very definitely looks like all Pro 5700 XT cards have the corner issue and the white lines. The white lines tend to go away when switching to a non-default resolution. The corner issue persists in any non-default resolution.
Yes, it is already established in this thread that the problem does not occur when a secondary screen is attached. So logically you don't experience the problem. It only occurs without a secondary screen and with the display resolution in default mode and with a Pro 5700 XT. If you test that you more than likely will have the glitch also.
You're potentially onto something here. Earlier in this thread I had a similar thought and asked myself the basic troubleshooting question: what's different with these cards?
AGDCDiagnose tells me that DSC is not used for the internal display. Or, more specifically, that it is off. One would expect this as why on earth would Apple make the display 30bit and TrueTone but then cripple the whole setup with chroma subsampling?
But: is there an issue with the hardware or drivers that requires DSC to be turned on for a display and that eliminates a bug of some sort with the non-DSC internal screen.
Yesterday I noticed the glitches several times, mostly while running the migration assistant but also on the normal desktop. Today I worked several hours with my new iMac and I have not seen any glitch. Maybe because most of the time I had bright windows, which would make the glitches harder to notice.
Secondary monitor. There you go... Disconnect it, set your screen to the default resolution and glitch away. Have fun!!!
I've been a lurker of this thread for the past couple weeks, checking everyday to see if a fix has come in. I'm ready to pull the trigger and purchase, but not until Apple fixes this.
Hmm, I wasn't able to make this work at 2667. 4gb 4gb 32gb empty did however
Me2. I need a 2nd iMac, but I’m holding off on the second computer until a fix comes. Fool me once......
Why hadn't this site, macrumors.com, had an article on its News Feed to draw this problem to the attention of its wider audience? I am sure many potential buyers are unaware of this problem in the 5700 XT iMac, and need to know of this problem. I guess the answer to that question could be the same for so many other issues we see in the worldwide media.
Well, based on what you claim...this would have to be a hardware fault.
If it were a software issue (as suspected), then everyone will have the same issue due to us all having the same software.
If however, you have a "copy that as of yet does not have any flashing lines", that would have to mean this is a hardware issue, and you somehow got lucky.
Personally, I agree with Auralux, and think you just haven't seen the issue occur as yet.
That’s right. I made me member on as many Facebook Mac related groups as possible to address the problem . Replies are all sort off “Which glitch ? “ Doesn’t that say enough. No , apart of our community the world is not aware from the problem.
Exactly right. Most people will never know about the glitch until they have already purchased their iMac. And only upon trying to research and troubleshoot the problem will they find this forum. We need a mainstream article written to bring awareness (which might also help pressure Apple into fixing the issue).
I’ve been informed my last reply was in poor taste, I’m sorry.
Thank-you for trying to spread information on the graphics card issue.