Interesting. I messed around with resizable BAR, but I didn’t do that in conjunction with IOMMU. I’ll give that a bash.
You contributed more than you can ever know 
1 Like
Hey mods cap me already
1 Like
And here’s a new thing. First cold boot after being off for a few hours (I made another post that explains that).
Issue disappeared on reboot, or at least it has disappeared for now.
I’m going to try the suggestions above now to see if I can get the second card working.
Thanks @wendell, I tested with resizable BAR on auto (no enabled) and IOMMU disabled, sadly with the same result. Appreciate your reply.
1 Like
Since no mentioning of it I guess the following didn’t help?
If not, did USB Flashback actually work (blinking LED on motherboard for a few minutes)? As mentioned I’ve experienced a motherboard state where USB flashback itself could no longer be triggered.
1 Like
That’s on the list to attempt later today. Apologies for taking so long to get to it.
No worries, I myself should be testing much stuff on personal systems that have no obvious causes like defective components and sometimes am one step before throwing everything out and becoming a technology-abstinent ascetic.
*Edit: grammar correction.
1 Like
That’s more appealing than it should be
Regarding your photo of the display only showing noise, this seems like an HDCP handshake failure, can be caused by:
- GPU drivers
- GPU ports (loose plugs, dirt, dust, corrosion)
- defective display cables
- Display firmware hang-ups
etc.
Would power-cycle everything (disconnect display from PSU or AC power, let it discharge internally for a few minutes) and re-plug any cables a few times to get maybe better contacts.
It is most likely one of these, the system is currently getting moved and fiddled with a fair bit as I work through some stuff.
It usually lives in a separate room where cables passed through a wall to limit office noise and heat. The current cables are far from idea, but it saves me needing to fish cables back through a wall.
If it happens again I’ll be sure to try that. Thanks.
Are you using any adapters? What cables are you using?
No adapters. These are short DisplayPort cables that came with a couple of Dell monitors, no idea who makes them, no branding.
The permanent cables (running through the wall) are Lindy Chromo, never had any issue with those, even at 165hz.
Depending on the configuration I use active adapters, yes (to get HDMI out of a GPU that only has DP out, for example).
Or I would like to use an active adapter to adapt DP 1.4-to-HDMI-2.1 but every (Club 3D) adapter that currently exists does not work properly for me, no matter how new the GPU is (tested up to a 3090, just to check if it is finally possible).
(But I get that you probably meant the OP of this thread)
But on the other hand, if someone sees this reply by chance while looking for a specific cable recommendation, the only DisplayPort 1.4 cables that worked without any issues for me (also using Displays with DP 1.4 like a Dell UP2718Q, that needs DP 1.4 to function fully since it does not have an HDMI 2.1 input) are the ones from StarTech:
Ok, finally got to test all of this. Sadly no real change.
Flashback worked, I took the system back to F11 and did some testing. I then went back to F12i, both could be confirmed from the on-screen bios version with only one Titan in the system.
Oddly post flash (to F11 and subsequently back to F12i), Fast-boot is now occasionally having issues.
I retested with IOMMU on and off. Resizable BAR on and off. Above 4G decoding on.
I’ve also tried to clear the CMOS entire as per your suggestion, other than needing to re-setup RAID, it made no difference on this system.
No change, still no boot with two Titans installed. Appreciate the suggestions.
For my own peace of mind: Also tried taking the CPU physically out of the socket while the CMOS battery had been removed?
I’m not suggesting this to be a jerk wanting you to potentially waste thermal paste, there definitely is a non-volatile settings memory on AM4 systems and any questions I asked about these to various parties in the past have so far been ignored.
A past thread addressing this:
1 Like
I did not. I’ve ordered more thermal paste. I’ll try this and report back when it arrives.
Sorry, I should have mentioned that I didn’t remove the CPU.
1 Like
This just keeps getting stranger and stranger.
I wasn’t going crazy when I thought the system was booting into Windows.
Using remote management software I have been able to connect to it and “use” windows absent any video card.
It’s almost like slot 1 and 2 are getting completely disabled when bifurcated for these two cards.
That might be where the reseating of the CPU helps resolve this in some instances.
Thermal paste should be here next week so I can try that. It’s a real head-scratcher now. I could kinda live with it when I thought the board just wouldn’t post with them installed.
If the reseat doesn’t work I’m probably going to RMA the board.
This is in-line with my observations with the two Radeon VIIs on a different motherboard (ASRock X570 Taichi Razer Edition).
My lay-person gut says that the UEFI does not initialize the firmware of the GPUs due to some AGESA bug.
This results in the CPU not using/ignoring its PCIe lanes that go to the GPUs’ slots since during POST it couldn’t detect anyting there properly.
GPUs aren’t the only thing where I have observed this: My first issues had been with NVMe SSDs in PCIe bifurcation adapters where only one of two drives showed up and the other one was ignored.
Now what needs to be done is finding the cause of this bug.
1 Like
This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.