■■■■, I feel your pain. I RMAed a PNY 5090 (after loving their 4090) and then had to return that second card because it was also terrible. Far too loud, making very weird noises, running hot and randomly revving up for no reason). Also got an Aorus Master next. Love that card!
Strange that you are still getting that error with the older drivers. I guess you’re not overclocking or undervolting?
I’m now on 581.08 because I was trying to get rid of stutters that started with SU3. Working good for now, but haven’t tested enough, error could come back any time.
Have you tried 576.88? That one didn’t have the error for me either.
Thanks mate - Yeah, I had this exact issue non stop with my first 5090 which was an MSI. I returned that and got a PNY which worked without issue for several months until the 12V2X6 power cable melted.
Just trying to get the Aorus 5090 working now. No overclocking or undervolting (although may try to reduce clock speeds next). My test scenario is just to launch the C208 out of EGLL, head east at 8000 and see if it crashes.
So far drivers I’ve tried are
576.52 → Ended in the 0x887A0006 error after 1 hour
576.28 → Ended in the 0x887A0006 error after 30 minutes
581.08 → Barely lasted 20 minutes.
I’m doing a full shader removal and running DDU between drivers.
I think the most infuriating part is that this error has been talked about for years and the developers have failed to address it, or even discuss it.
Oh god you had the melted connector =(
Undervolting should reduce the risk.
576.52 and 576.88 are both rock solid for me, never got that error with them, ever. From 577.0 it always came back at some point after 5 to 30mins.
I’m sure it will with 581.08 too, hope not though. If it does I’m going back to 576.88 and hope the stutters stay away. (The deciding factor may actually have been deletion of the scenery indexes, which I did at the same time as upgrading the driver).
Can I ask what might seem like a silly question but I am not a computer knowledgeable person. Why would one overclock a 5090 if it is already the fastest GPU on the market? How does it benefit a flight or whatever the goal is one might be trying to achieve? Genuine question, I really do not know.
In any case, the 5090 is still too slow to drive high res VR headsets at their 100% resolution at acceptable frame rates (45+), especially without making a lot of concessions in the settings.
So if I could overclock a 5090 in such a way as to make an impact full difference, I would. Gains would be marginal at best though so it’s not worth it. 5090s benefit more from Undervolting.
Yeah I never actually thought it would happen to me, but sure enough! Still waiting for PNY to accept the warranty claim . Once I find a stable driver with this card I’ll look to undervolt it to try and reduce the risk. Guessing MSI Afterburner is the easiest way to go but happy to hear advice from anybody on this …
I’m away from the PC until next Friday so won’t be able to test other drivers, but I’m hoping either 576.88 or 572.?? works for me. I’ve got a Pimax Super coming this week or next and I really wanted to try it out in 2024!
I think its important those of us who are getting these errors (and its clearly quite a few) put the details here do others can test what may have worked/not worked. Its interesting that the older drivers are the ones that seem more reliable in this case…
Friendly reminder: Software like MSI Afterburner and GPU undervolting can sometimes cause stuttering and DXGI errors - including audio stutters. Speaking from personal experience, so I’d recommend using them with caution.
Odd - For me 581.08 resulted not in the usual 0X887A0006 crash dialog box, but rather in the game just freezing up. What was interested was that when it did this, task manager would still show MSFS as running (i.e. not “not responding” and I could see the GPU was still running at around 2600-2700 MHZ, like it was being asked to work. At the same time however, the power draw to the card dropped off as did its GPU workload down to 1%. It was like MSFS simply stopped sending the GPU data to process…..which I think is often what the 0X887A0006 error refers to anyway.
I also wondered whether that could be the cause. But removing the undervolt made 0 difference, I still got the errors (with specific drivers). Still a valid point you raise though.
In the case of 5090s, undervolting is quite important since it lowers the risk of connector melting. And costs virtually no performance or stability when done right (conservatively) , as I’m sure you know.
You’re right about undervolting, but it’s also important that the PSU has a proper connector – a single one, not three separate ones. And definitely avoid powering the GPU through adapters as that only adds risk. But that’s a whole different topic …
My PNY 5090 that melted was using a three month old Montech ATX 3.1 Titan 1200W Platinum PSU with the PSU supplied Montech 12V2X6 cable.
What I find interesting about this particular crash is that its so dependant on drivers/individual graphics cards. Like I said above, I had an MSI 5090 that would crash on every single driver released when running MSFS 2024. 2020 was fine, but 2024 would crash within 5 mins on each load. Multiple re-installs, including several full re-installs of Windows made no difference. I swapped it for a PNY 5090 and using the same drivers the PNY worked like a charm. Then the Gigabyte I’ve currently got didn’t work with those drivers. I
ts just mystifying made more frustrating by the fact that Asobo are silent on the issue.
I totally understand you, but it’s really best to stay away from third-party adapters and converters. PSU manufacturer should take care of that by providing the proper native cable. If you end up using adapters and something goes wrong, you’ll never be able to prove whether the fault was on PSU side, its cable or GPU. With the original cable that comes with PSU or GPU, it’s straightforward to show responsibility if anything happens. Some GPU manufacturers do include adapter cables, but those are mainly meant for older PSUs that don’t have the proper connector. Ideally, you want to avoid such PSUs altogether - PSU is not the piece of hardware to cut corners on.
Sorry for going a bit off-topic but I thought it was an important point to make.
. Once I find a stable driver with this card I’ll look to undervolt it to try and reduce the risk. Guessing MSI Afterburner is the easiest way to go but happy to hear advice from anybody on this …
