NVIDIA 790i SLI Data Corruption - Solved?

27 Comments

View All Comments

• akidd - Friday, June 6, 2008 - link

  Sorry - I am using stock clockspeeds - no overclocking.
  
  GPU clock (geometric domain) 600MHz
  GPU clock (shader domain) 1500MHz
  Ramdac 400MHz
  Memory: 1000MHz

• FXi - Friday, June 6, 2008 - link

  Well it's a mark of credit to Gary and the Anandtech team that they are going the extra mile to try and help the community test for more bugs. That's the community coming together in a good way to try and get to the root of the problem. At least if it they can determine what boards need to be RMA'd, they'll help frustrated folks find some resolution. (hopefully) So kudos on a good effort there.
  
  As for Nvidia's chipsets, one has to ask how many times you want to be fooled. As far back as I can recall there have been issues. Drive corruption is hardly new to this chipset. It's just frustrating that it went out the door with that problem, after years of experience telling them they should have been more careful. But there just have been issues with every single generation. How long do we, the community, even bother to try them just to hope they finally got it right?
  
  There's a reason they are hesitating to cross license Nehalem for SLI. They know, perfectly well, that their chipset business is either dead or it's going to have to take a HUGE step up in quality to stay in the game. And after years of trying, I have a guess that they know perfectly well they can't make a good chipset, so they are doing anything they can to protect it, preying on the ignorant customer for as long as they can. There is no way they could survive at the current quality levels. And frankly, after so many years of trying and failing, their chipset business should be long dead.
  

• ATWindsor - Friday, June 6, 2008 - link

  It's nice that they have fixed it, but this is not the first data-corruption-issue with nvidia-chipsets, they will have to show a good track-record for some years before I trust them again.

• Tanclearas - Monday, June 9, 2008 - link

  Nvidia claiming to work with Gary and the community is nothing new. You can look for my post to Gary's last article on nvidia corruption issues in the link below.
  
  http://www.anandtech.com/mb/showdoc.aspx?i=3279&am...">http://www.anandtech.com/mb/showdoc.aspx?i=3279&am...
  
  Gary cared about the issue far more than nvidia did. You can be sure that nvidia will try to make it look like they are working with the community, but their ultimate solution will be to tell you to buy a new board with their latest chipset.

• Inkjammer - Friday, June 6, 2008 - link

  Frankly, I've had incredible amounts of trouble with my eVGA 780i... lots of bugs, quirks, other problems. The 780i's notorious video corruption problem was FINALLY fixed, but I still have other problems.
  
  I'm really starting to doubt the QC work there. I know new tech is never 100% flawless, but some of the flaws that they have have been rather high profile.

• cngn - Friday, June 6, 2008 - link

  Bought and EVGA 790i Board, had enumerable problems with just keeping the system running. I guess it never got it running long enough ever corrupt a disk.
  
  I use the PC to Fold (http://folding.stanford.edu/)">http://folding.stanford.edu/), but the moment the CPU went to 100% usage the CPU Vcore dropped so low the machine would just hang.
  
  Tried upping the Voltage on the CPU the Memory even the FSB and NB but to no avail I just couldn't get the machine to run for more than an hour, without it hanging
  
  Finally gave up on all the crap I was doing and have now bought myself an ASUS P5e3 Premium (X48) and it runs 100% 24/7 without one issue at all.
  
  I'm RAM'ing the board back to them but to be frank I'm not holding my breath on the next board working any better.
  
  Nvidia needs to climb back into it's little hole making good GPUs and leave the other more complicated stuff to people who know what they're doing I.E., Intel. The 680i, 780i and now the 790i, all started out with the same problems, data corruption and the VDrop and VDroop problems, if they can't get it right this far in the game.. I wouldn't be buying any shares in them.
  
  And for God's sake release SLI licensing to Intel, stupid asses, if they think they're being smart by not doing so, I'd say they're cutting off they're noses to spite their faces.
  
  I for one would rather use AMD's Crossfire tech now since I'm forced to use a chipset that works.

• Tig3RStylus - Monday, June 30, 2008 - link

  I am afraid i have to agree, quite an experienced overclocker, but never really got a chance with the 680i. I was unfortunate enough to get the double whammy 680i and the ASUS version. When your board dont work, you want a warranty... or at least somebody who responds.. mines still in its box, its been in it since November 3rd 2007 when id exhausted 2 months of trying to work at stock speeds. They never got back to me, ive lost all will to continue chasing and put it down to a bad investment. PS if anybody wants a striker extreme as frisbee or stress toy, let me know.
  
  It had the dreaded CPU Init problem, it also destroyed 2 sticks of DDR2-1066 which was quite expensive when it was purchased, fortunately Corsair have excellent warranty and no quibble replaced but commented about the 680 being a ram killer... probably from their warranty costs hitting the roof. Also somebody else commented, forget 4gb of ram or more.. The board wiped out my Data 4 times before i put it back in the box and decided it was a good idea to stick with tried and tested Intel boards.
  
  Ive been running an X38 since November and now have a Rampage Formula X48 running stably with a tidy overlclock. Im sticking with this until nehalem upgrade, this is plug and play for initial setup at stock speeds and completely stable overclocker if you want to thrash it.
  
  I wont be believing the SLI marketing machine until its on an intel board... Do and your in for a whole world of trouble..I rest my case.