5080 Suprim on MEG ACE x670e

[frozen_12157b02de]

I recently bought a 2nd hand 5080. I tested it in my older AM4 system at the time of purchase to verify it worked. No issues at the time. I've had it sitting in my office here for a few weeks and finally decided to install it in my workstation, daily system. Not a fun experience.

The system had a 4080 Super in it. The drivers were up to date within a few weeks to a month. I figured it would be as simple as swapping hardware. Upon doing this, and booting back into windows, I immediately started noticing issues. The start menu was slow to respond, the displays would turn off, as if having lost signal, and then come back on after a few seconds, there maybe have been some flat out restarts too. So much happened, its hard to remember it all. I tried installing a slightly older studio driver, doing a clean install. The problem persisted. At some point i tried putting the 4080 back in, and I believe the problem STILL persisted, which was very worrying. At some point i noticed in the Device Manager that there was a "microsoft basic display adapter" listed, with the orange warning triangle. I know the ACE has integrated graphics. At some point i tried disabling that in the bios, but it made no difference, though it did remove that listing from Device Manager.

I thought id use the DDU utility from starting in safe mode, but i never got that far. One of the restarts or freezes or crashes caused windows to go corrupt and it would not start.

I then had to turn my attention to getting windows running by restoring from a system image, and that also became a huge problem because it seems the ACE likes to just do weird stuff that i cant predict. No matter what i tried i was unable to use the USB recovery stick to get into the windows recovery environment. I finally ended up doing a CMOS reset, using a brand new SSD, removing the old windows ssd, and was somehow finally able to use the USB to boot to windows install. Once i got windows installed i installed the same almost up to date drivers as mentioned above, put the 5080 back in, and the problem persisted. So this is now a completely fresh install of windows on a completely fresh SSD.

So now im realy getting upset/worried. I put the 4080 back in, and I was finally able to put the original windows ssd back in, and get the recovery environment to actually boot up, and restored the windows image from january i had, for just such a situation. Im back up and running now, and hopefully will be able to get on with work tomorow morning without any issues, though now I have that what-if fear.

I saw 1 reddit thread briefly on my phone where someone was complaining about a potential hardware conflict... i dont know if thats valid? The problem didnt seem to exist in bios, so it seems to be windows only. However 2 different driver packages didnt change the results. Also strange that this 5080 worked fine (from what i could briefly tell) in the old AM4 system. So its seems to be specific to this hardware combination.

I skimmed through another reddit through where LOTS of people were reporting issues with the 50 series and various display artifacts. Some also saying they were having the issue with 40 series cards. This 4080 has been, as far as i can tell, mostly stable, though now that i read these other threads, i have noticed for many months now that, occasionally, when waking from sleep, i will get a similar problem where the display will seem sluggish. everything is choppy. menus are slow to respond. only fix is a restart, and then it works fine for a few days. this behaviour was actually what prompted me to look into a new 50 card, thinking perhaps the 4080 was showing early signs of a problem. But now im starting to lean towards there might just be something more fundamentally wrong with the drivers for both card series.

anyway, if you' read this far, thank you. Im hoping someone has some insight, though at this point, im not hopefull. I also dont have the luxury to be on a flaky system when im in the middle of a large project, so i need to keep going with the 4080 for now, and wont be able to do any testing with the 5080 back in this system until some time in june. But id welcome any suggestions anyone might have. Again, ill reiterate that things were working fine, and NOTHING CHANGED other than the actual gpu. Removed 4080, installed 5080, chaos ensued.

A few final points that might be relevant...
-7950x
- bios 7d69v19 (approx sept '23. quite out of date, but working fine)
-win10 22h2 (same as the am4 system originally tested on)
-1000w psu, w/ proper 12vhpwr. quadruple checked the cable was properly seated
-using 2 Dell monitors via displayport (did not check if the problem persisted via HDMI).
-both monitors are 60hz only
-using studio driver, so GSync is not even an option to turn off.
-all nvidia settings default

thanks.

 

[plutomate]

Regardless the system works fine before or not, you should always update BIOS first especially for problem like this (compatibility?)

https://download.msi.com/bos_exe/mb/7D69v1O8.zip

 

[Svet]

Moved topic to correct area.

-1000w psu, w/ proper 12vhpwr. quadruple checked the cable was properly seated

Which brand and model it is?

 

[frozen_12157b02de]

Moved topic to correct area.



Which brand and model it is?

thanks for the replies, and moving to the proper area. I wasnt sure if it belonged in the GPU area.

the PSU is an FSP Hydro G Pro. Not my ideal choice, but at the time (about a year ago), it was the only thing I could find with a proper 12vhpwr connector when I upgraded to the 4080. I didnt want to risk using the pci-e pigtail adapter on my Seasonic PSU i had previously.

Regarding BIOS, yes, i know im long overdue for an update. But i've been burned on bios updates failing a couple times long ago, and its always made me hesitant since. I usually take the 'if it aint broke dont fix it stance' when it comes to bios updates. Up until the 5080 went in, it wasnt broken, which is why i never updated it. Once i have the time to do a more thorough test with the 5080 when im not in the middle of a project, i will update the bios and see if that does the trick. But again, my thinking was if the 5080 worked on an AM4 with MUCH older bios, then the fact that my current one is a couple years out of date should still be better than whatever that AM4 is running.

 

[plutomate]

Then can you try lock Pcie gen speed to gen 3 in BIOS?

 

[Svet]

Regarding BIOS, yes, i know im long overdue for an update. But i've been burned on bios updates failing a couple times long ago, and its always made me hesitant since. I usually take the 'if it aint broke dont fix it stance' when it comes to bios updates. Up until the 5080 went in, it wasnt broken, which is why i never updated it.

It's a time to update it, don't afraid of bad flash as you have (your board) a Flash Button.
And you really have issue now and reason for update it.

So go ahead and flash it, then see if that helps.

 

[frozen_12157b02de]

Thanks again for the suggestions. As i said above, i cant try this again right now. I need to wait a bit until im between projects before i can risk having a whole windows crash again. But, that said, when i do have the time, i will update bios and see if that helps. Most likely i will try again in this system some time in June.

As far as the PCI gen switch, im assuming that suggestion was based on NOT wanting to update the bios? As a work around?

I decided to try the 5080 again in the AM4 system yesterday afternoon. I got some interesting results. First off, ill say that i was incorrect about that system being older. I dont use that system very often and it completely slipped my mind that, about a year ago, the MEG ACE x570 which HAD been in that system had died, and it was replaced with an MPG board. So its actually probably more up-to-date than the MEG ACE in this system, with a more recent bios, since that was just a year ago or so.

Anyway, i had been speaking with someone else and they suggested a few things which I wanted to try. I installed the 5080 and booted it with only 1 DisplayPort connected. I brought the whole system to my office so it was using my normal everyday monitors. The initial tested id done with the AM4 system had actually used an HDMI monitor, and it was suggested that maybe the displayports where the cause, so I wanted to try and rule that out. Upon booting into windows, at 1st, all seemed well with 1 displayport monitor plugged in. I downloaded a recent version of furmark just to give it a bit of a stress test for a minute, and it ran fine with no issues that I could pick up on. Everything was still responsive.

I then plugged in the 2nd monitor, and thats when i started to get some of the same odd behavior as id experienced in this system. The mouse started lagging and stuttering, menus would freeze or not open. I unplugged the 2nd monitor, and the problem stopped. So i was thinking maybe its an issue with running 2 monitors at once, or using 2 displayports at once, or 1 of the displayport cables was bad. So here i get a bit fuzzy on what exactly I did, but i believe I tried an HDMI cable into the 2nd monitor, and then also tried a DIFFERENT displayport cable, as well as trying a different port on the back for the displayport cable, thinking maybe the port itself was somehow the issue. So after a bit of cable shuffling, i ended up with the original displayport cables plugged into the same ports on the back, and the problem started to go away. Im not sure why. This whole process only took about 2 to 3 minutes. Im not sure if windows needed to "figure out the hardware change" and was internally doing something that took time to smooth out? Maybe that was happening on this system and i just didnt give it a chance? Though it seems odd that the problem appeared right when i added the 2nd monitor, and stopped when i removed it.... even if it was still internally doing something.

After the mouse stopped lagging, things seemed to behave normally. I didnt do much more testing at that point. I figured the exercise showed me that maybe i just needed to do the same thing in this system and see if the issue self-corrected as it did with the AM4 test.

Anyway, im not sure if that is indicative of anything or what. But i just thought id share that experiment incase it means anything.

Thanks again for the continued input on this.