Issues with 2nd GTX 1080 Ti card...

smackthis527

Not sure where else to put this, so I apologize if this isn't the right spot. I have two 1080 Ti cards in a rendering machine that has been going just fine...until recently. For quite a few years now I have been able to use both cards to render. Obviously not for the Vram, but just for the power to make render speed faster. I was having an issue lately cause I would render one scene just fine and then the very next one would start but crash my entire computer. I've done some tests. I went to my top card that was manufactured by ASUS, and that did a render load of 38 images fine, though it took like a day and a half. I tried the same test on the 2nd card manufactured by ZOLTAC and it reproduced the crash after just one image. I'm running Daz 4.24 right now. I don't want to upgrade to the newer Daz for fear it'll bake my 1080 cards since I can no longer update them. I tried changing the final updates of the drives, no luck. Still crashing. I ran the card through OCCT and no errors at all came up. So I'm completely lost on what to do. Has anyone else experienced this and come up with a possible fix? Or anyone have any kinda idea on how to handle this that doesn't have me going out and dropping $4k on a new card? :-p I know someone's gonna say it, haha. Seriously though, any help would be welcome. Thank you :)

mikael-aronsson

Obvious question, have you checked the fan's on the "bad" card ? checked temperatures on both cards when they render, just so there is no overheating issue.

If it worked before and does not work now, either something is broken on the card or something has been updated, there is not much else that could happen.

richardandtracy

Or possibly is the second card getting clogged with dust and overheating? Does the same thing happen to the same card if the two are swapped in position (motherboard power/data to card issue may need considering)? It's possible something is failing on the second card and only shows up when hot, if it can't cope when warmed up by the first render.

Difficult one

Richard

smackthis527

richardandtracy said:

Or possibly is the second card getting clogged with dust and overheating? Does the same thing happen to the same card if the two are swapped in position (motherboard power/data to card issue may need considering)? It's possible something is failing on the second card and only shows up when hot, if it can't cope when warmed up by the first render.

Difficult one

Richard

Changing the cards placement is a good idea. I had checked and the fans are working and there is no dust, but I didn't see if maybe It's that slot and something not showing up til it's hot. I'll give that a try and let you know how I make out. 

smackthis527

mikael-aronsson said:

Obvious question, have you checked the fan's on the "bad" card ? checked temperatures on both cards when they render, just so there is no overheating issue.

If it worked before and does not work now, either something is broken on the card or something has been updated, there is not much else that could happen.

I checked all the temps and tested the card in OCCT. Everything checked out fine for the suspect card. Someone else suggested I switch the cards to see if the suspect card is still an in issue in the other spot. Maybe something is going on there in that spot when it gets hot. I also had double checked everything with Daz and my cards, nothing had been updated. The dump file indicates when the second card is done a render, it stalls out and that's what winds up crashing my computer. So I want to give a shot at sliding the spots around and see if that does anything. 

davidwski_16294691f0

smackthis527 said:

mikael-aronsson said:

Obvious question, have you checked the fan's on the "bad" card ? checked temperatures on both cards when they render, just so there is no overheating issue.

If it worked before and does not work now, either something is broken on the card or something has been updated, there is not much else that could happen.

I checked all the temps and tested the card in OCCT. Everything checked out fine for the suspect card. Someone else suggested I switch the cards to see if the suspect card is still an in issue in the other spot. Maybe something is going on there in that spot when it gets hot. I also had double checked everything with Daz and my cards, nothing had been updated. The dump file indicates when the second card is done a render, it stalls out and that's what winds up crashing my computer. So I want to give a shot at sliding the spots around and see if that does anything. 

How old are the cards. You mentioned several years. It's also possible that the heat sink paste has dryed out and no longer performing well. It's relatively inexpensive to get checked, or even replaced. It should be replaced periodically anyway.