Crossfire for over 2 years. Suddenly one Direct CUII 7970 TOP is not recognized...

[SplitSecond]

Hey everyone,

So I have been running crossfire for over 2 years and now suddenly one card is not recognized anymore.
I have tried switching PCIe lanes, running only the "faulty" card, updating drivers, looking in MSI afterburner, catalyst control center,
but nowhere does it show my second card anymore.

I never had any trouble with the card, it also never has been OC. Is it dead...?

Please help!

[ChiefZeke]

I'm assuming the first thing you did was to simply swap the two cards around - #1 card in #2's slot and vice versa? If that worked OK and the problem continued to exist then the cards would not appear to be bad - possibility would then be the motherboard PCI slot itself.

[SplitSecond]

I have tried swapping slots, that didn't work. In addition, I tried running the PC with only one graphics card at a time, with one it ran without any difficulties, with the other it didn't show anything at all. That's why I guess the card might be defect 😞

Asus RMA in The Netherlands also doesn't recognize my serial number, which is not only strange but also very inconvenient since I really want to have a replacement for this card (because of crossfire). As a student buying two new cards is rather unrealistic, I even saved up for these since they were cherry picked 😕 Besides specs this one also fits my custom case exactly as I want.

Hope Asus ROG or someone else has a solution for me that I can try!

[Korth]

Maybe your CrossFire bridge cable is damaged?

[SplitSecond]

I replaced it with another one, but didn't help. Thanks for the suggestion tho! This also wouldn't explain why the test with the single card wouldn't show something on screen 😕

Please help me with suggestions. I can't miss my crossfire!

[Korth]

So one card works in either slot, the other card doesn't work in either slot?

Most-to-least likely causes of the fault:
- your firmware/configuration (maybe try updating to latest-greatest BIOS version, clear CMOS, replace mobo battery, etc?)
- your power supply is starting to die (no PSU lasts forever, easy to prove/disprove by testing with a robust new PSU)
- a PCIe connection (you might be able to blow out dust/etc from the mobo slot or use a contact cleaner on the card)
- a GPU card (or maybe just the cooler or some other component on the card)
- your motherboard (electrical components do eventually fail, especially if frequently overclocked or overheated)

I'm thinking the mobo battery is the culprit (a CR2032 coin cell typically lasts about 18-30 months), since you describe your system as being about 2 years old and working fine until the mysterious CF failure.
BIOS settings and SSD data (including your OS system files) can be also corrupted by sudden power loss events.
Catalyst drivers are notorious for stubbornly insisting on ninja self-updating ... and sometimes newer is not always better.

Doesn't hurt to change the battery (if it's not dying now then it probably will be soon, and 3 for $1 lol).
Manually check all relevant BIOS settings.
If you ran AI Suite or some other autoconfiguring software to get stable system settings before then run the exact same version again.
Maybe try a rollback to a previous driver configuration, Windows Restore Point, etc.
Sometimes PCIe cards (and DIMMs) need to be reseated a few times to work correctly, not so much intermittent contact as just weird and quirky.
Visually inspect your cards and mobo for signs of component damage (broken solder points, cracked PCB traces, leaky electrolytics, burnt resistors, etc) and check all the ICs to confirm chip creep hasn't pushed anything out of its socket.
Look for signs of galvanic fretting in all connectors and sockets (a type of corrosion caused when dissimilar metals are in forced contact).

That's all my ideas, lol. Good luck.

[SplitSecond]

I found the problem I think...

Recently my other card also became a brick. Both cards dying so fast after each other seemed kind of strange to me. Given that I have no warranty left I decided to disassemble both of my cards. In each card I found 2 out of 6 heat pipes to have melted to the point where liquid coolant could escape and drop on the card itself. Needless to say having no effective cooling is kind of a death sentence for graphics cards... Interesting to note is that both cards have operated in a well ventilated case and have always run in default mode on a Rampage Extreme IV. Seems like a manufacturer malfunction in my eyes.