I had wondered if it was newer released card vs original with differing bios myself. I know on my card I have the latest offered bios. Still doesn't explain why the older driver loaded and the 2019 drivers don't . Awesome if you at least got yours going. Good Job!
FYI, my friend that is also running a 280X DCII Top without a problem on latest drivers is using that exact same vbios (for anyone with a R9280X-DC2T-3GD5, this bios update is available directly on its asus support page, listed as Version 1.0 2014/01/26).
Edit: Btw, that asus bios only has vram tables for Hynix H5GQ2H24AFR, so please don't atempt to cross-flash it on a MSI Lightning, because the Hynx models use H5GQ2H24MFR (as most HD7970s), not to mention that those are both different custom PCB boards, not reference designs.
Viniabear and cesarin's current 7970 Lightning bios (same) is a very old 015.013.000.010 version prior to UEFI support. Probably the latest was 015.013.000.011.000000 as listed on techpowrup DB, but again, you lose nothing in asking msi support for the latest bios if you want to risk the chance of being more compatible with the latest drivers. With some luck, there could even be a latter rev. with added UEFI support (there was such update for the MSI 7870 I'm using in one PC).
Still, this is something that AMD should look at, as it seems an incompatibility recently introduced on driver side. Many old 7970s models might not have the same luck of having a vbios update, unlike later 7970 GHz and 280X models.
Edit2: Unfortunately no similar update is listed for ataman's R9280X-DC2T-3GD5V2, but you might have some luck by contacting Asus directly. On a side note, I've had a similar triple slot 7970 DCII Top (with Hynix MFR). If this card is like its predecessor, it requires two model specific vbios because of its video outputs (controlled by bios switch).
Edit3: cesarin I forgot to ask the most basic question, what happens if you DDU and then install one of these drivers directly via Device Manager (by pointing to c:\AMD\<driver version>\Packages\Drivers\Display\WT6A_INF? Per your log, this is just to check if the BSOD problem is with the driver or something the installer tries to do.
You've already left clear that 220.127.116.11 can install without a problem via their setup. My question was if you can even install the 19.3.1 driver-only directly via device manager, to check if the BSOD problem is right within the driver or just the installer routines. If it's possible to install the driver manually, you can then add the CCC2 (via ccc2_install.exe), vulkan and hdaudio driver too (provided any of these isn't the culprit), and at least have a workaround until AMD fixes their stuff - OTOH, if the solo driver fails, thing's wont be so easy.
Please understand that the installer has drivers for many different configs and has to do some check routines that are really meant to fail if no suited hardware is found -- some errors are to be expected. As an example, that amdhdaudbus.msi isn't meant for your system (the correct will be AtihdWT6.msi or its inf for the AMD High Definition Audio Device in hdmi). In the same way, amdacpbus.msi (another audio controller) also isn't meant for your hardware.
I notice in the error logs "risk mitigation" messages. This may be some internal AMD level driver stuff or it might be coming from windows secure platform. Your latest screen grabbed the GPUz info has an unchecked uefi box. The two are potentially related at a windows level and at the motherboard bios level. I opine it something to do with windows since your getting a blue screen error as well as a log output ( but I don't rule out both parties here ).
Have you tried disabling the compatibility support modul ( CSM ) in your bios? This is located in the Boot section of your bios, scrolled down a bit in the options. If it is set to Auto, set it to disable to disable it ( if you so choose ). Page § 3-53 of you motherboard manual. If you have already done this or tried this, please know it is only my best intentions. If you are able boot into windows, you may have better luck if not different error messages at least.
Auto is not a good setting for the CSM.
Obviously, I will understand if you do not want to try this alone, you have described installation horrors regarding not being able to properly DDU or otherwise after attempting and the resulting failing of installation.
Good day for everyone, I think the problem is not from the BIOS of the video card, rather from some process that is doing badly the installer package of AMD 19.3.1. and similar in this particular model, since it compares the two INSTALLER.LOG files of each of the versions, the one that serves (18.104.22.168) and the one that does not, (19.3.1) and can see where it does not work the errors that it gives at the end, with which the process does not end and the other capture that if it is installed correctly.
¿Any engineer who can interpret these Installer.Log to help identify where the fault is?
Installer.Log File of 19.3.1
Although, as mentioned earlier, I have been able to successfully install the 19.3.1 multiple times on a Gigabyte HD 7970, we continue to investigate the issue.
I will update this thread when we have more information.
That's an interesting thought. In my case I don't even have a UEFI bios. Mine is an AM3 with a Phenom II. It's great to have some real constructive ideas on where the conflict can be happening. I just hope these ideas are passed to engineering. Hopefully someone isolates this enough to truly pinpoint the issue. I'm sure there has to be a common denominator in all this causing issue, but none of this changes the fact that the cards did and still do work with the older drivers. So no it isn't a bad card or other hardware being bad.
Thanks ray! I posted this and saw you just posted your reply. We very much appreciate you guys looking into this further.
@adamdz unfortunately so far the last driver that worked with this card is 22.214.171.124, are reviewing in engineering what is happening with the driver package to achieve a solution, as it is confirmed that it is a failure in the drivers with this specific model of cards. We hope that news give us in the engineering department about what they have detected and a possible solution to our problem. They have already passed all the tests and are investigating the case.
Thank you for your contribution as it is confirmed more and more that it is not an isolated case but it is a general problem with the HD 7970.