It looks like any post Fermi series is subject to the issue.....a refurb 580 would be a major downgrade from a pair of 570's....so that's a no go...I don't like it....but at the end of the day, I'm not going to sacrifice gaming performance because I have to shut/down/hibernate instead of going in sleep mode.
 
  Also when I sent the cards in I told them I wouldn't be opposed to and actually would prefer a single card solution of equal performance, they cant do that because each card is covered by its own warranty, even though they were purchased together , direct from EVGA with sequential serial numbers....policy is policy....

protivakid
 
 
  
What a crappy company...

 
   I wouldn't go that far, and the policy makes sense to a large degree....I'm sure the number of multiple card failures at the same time would see a huge increase if standard policy included an option for 2 cards being RMA'ed for a single card of equal or greater combined performance.
 
   Also my cards didn't die at the same time, one crapped out months ago, did a driver update and at first when I enabled SLI (as you always have to do when upgrading drivers) everything would go to hell, only one monitor would display, 3D wouldn't work, 800x600 was the only res option etc....shortly after that the option to SLI just disappeared, both cards showed up in device manager and system info programs like that in 3D mark, but in EVGA PX and any status monitoring programs one card was missing, but aside from that things functioned normally after one card disappeared . All that happened in late December/early January. Didn't have time to dig into the system for a good while, so I just ran off the good card and left the bad one in till I had time root around and figure out what was wrong.
 
   Months pass and I was finally able to work on it, pulled both cards and put what I thought was the offending card the in, as one was working fine still. The card I put in had major artifacting all over the place even at post and would blue screen in short order....that was easy, had things to do shortly so I called and started an RMA before putting the good card back in as I was pressed for time. Plans changed and suddenly I had several more hours free so I put the "good" card in and ....NOTHING....no post, no monitor activation...the fan would spin but that's was the only sign of life...put an old 460 in and device manager saw a card was in the slot, but that was as far as went....turns out that was the "bad" card all along....and the "good" card which was working just fine until 10 minutes before I pulled it suddenly was far from....that's how I ended up doing a double RMA....not sure how things would have played out if I did things differently, like pulling the bad card when the problem first happened.....also just now wondering if one or both problems might be related to S3 sleep...almost 4 years of up to a dozen naps and wake ups a day add up to a butt ton of possible intense voltage spikes.....its as good a theory as any.

iceman0124For those saying it has nothing to do with the cards.....I cant follow that logic, I got this motherboard and CPU in 2010, and have run many NVIDIA cards across a fair span of time (see above)....and the most recent cards in my case (660 ti's), made years after the x58 chipset was in its heyday ....and its "all" the motherboards fault? Cards before the mobo was made work, cards after the mobo was made work...these dont for some reason, they do sometimes but that's not really good enough...
 
  If anyone has a fix sans insomnia or hibernation please let me know.

It has as much to do with the cards, as it would be ram itself if you put two s ticks of dram in a motherboard that doesn't support them correctly (never seen those bios updates with "improves dram stability"?)
 
The issue is the bios on the motherboard, not the first time a mainboard has required a bios update to support an add-in card, won't be the last.
 
...atleast the machine works otherwise...

Considering its not an nvidia side issue (its a phoenix bios side issue due to hardware specifics of the nvidia card) thats unlikely.

squall-leonhart
Considering its not an nvidia side issue (its a phoenix bios side issue due to hardware specifics of the nvidia card) thats unlikely.

At one time Nvidia cards worked with this BIOS and then during the 600 series of cards something changed in the nvidia hardware. Hoping that whatever is preventing this from working changes again down the road and by chance works with X58. I understand this is all by chance and am not holding my breath.
 
Has anyone tried a new Radeon card to see if those have sleep issues?

squall-leonhart
At one time nvidia bios's weren't so large.
 
The issue is the phoenix bios.

Is that's what's causing it? The NV BIOS is too big for the Phoenix based board?

Mr. DOS

protivakidHas anyone tried a new Radeon card to see if those have sleep issues?

Well, my “solution” has been to swap out my GTX 660 Ti for a Radeon HD 6970, so...
 
I haven't tried any AMD Radeon Rx 200-series cards. A friend of mine has a couple; I might try to borrow one and see how it goes. I strongly suspect it'll behave flawlessly, though.

Very curious to see how this does for you. PLEASE let us know. And I'll assume with the 660 Ti you had the same sleep issue we all did?

I have the same issue with a 960 gtx.
I just had a talk to EVGA Customer Service. There will be NO fix for this. Buy a more recent product...they said.
 
Cheers!

Sorry,
I thought the topic was about EVGA x58 boards, or EVGA boards with Phoenix BIOS. Everyone is complaining about S3 resume not functioning regardless of BIOS settings/OS refresh. I landed on this topic searching for answers for my weird issue (EVGA x58 SLI board; GTX 960; S3 resume not working). Official answer was there will be no fix for any x58 boards. I'll return my 960 and go for an ATI 280x.
 
I think this info will be usefull for anyone having a similar issue.
 
Cheers!
 
UPDATE : I returned the 960GTX and got a 290x. Everything works as expected. Only downside with the 290x is that it consumes almost 3 times as much power, compared to the 960gtx. Out!

squall-leonhart
At one time nvidia bios's weren't so large.
 
The issue is the phoenix bios.

I am not sure how much the VBIOS size matters. The GTX classified 580 I run now has a VBIOS size of 81KB according to the site below. The Galaxy GTX 670 I tried when I first encountered this problem has a VBIOS size of 96KB which isn't much larger than the 580. You think that extra 17KB makes all the difference?
 
Also the high end AMD cards at the time X58 was popular in 2010 +/- 2 yrs such as the 6970 has a VBIOS size of 128KB which is larger than the Nvidia 670 yet no one reported having problems with those.
 
Would love to know the exact reason of incompatibility.
 
580: https://www.techpowerup.com/vgabios/112644/evga-gtx580-3072-111125
670: https://www.techpowerup.com/vgabios/127272/galaxy-gtx670-4096-120703
6970: https://www.techpowerup.com/vgabios/88328/ati-hd6970-2048-101124
 

Surely there must be a way to remedy this -- or at least determine a cause for why it can't hold a sleep state. I've had my x58 758 sleeping nicely for years, however, in these past 9+ months, it's suddenly having trouble doing so, with not much changed. Any ideas?