Video board HIS-6990 lasted only 3 days of mining...

[ThiagoCMC]

Friends,

 I got almost 2 Bitcoins mining with my brand new HD6990 and on the third day, GUP1 stopped... When loading Xorg with ATI drivers, the computer freezes... I changed my motherboard/CPU/RAM and nothing, hangs too...

 The miner at GPU0 still works, but GPU1 gives the mere 0.1MHash and freezes the machine in seconds... I can only activate the environment using the VESA/Generic driver of Xorg... If I try to load the ATI drivers, it freezes immediatly.

 The power source is ideal, 750W... Had not done overclocking, just a test to see if the commands worked, cooler at 100% all the time ... The temperature was at peaks of 92 degrees during the day and 89.50 º C at night... So I lowered the clock up to less than 880 during the day time... Incidentally, the machine was installed in a small data center of a company, with air conditioning and everything else...

 So here is the hint... Very careful with your video boards! I'm more than sad... God willing, they will change my board in warranty... For after all, I did nothing wrong... Or did I?!

 Did I do something wrong?!

Abx!
Thiago

[1715363622]

1715363622

[xunker]

92C is really too high for sustained periods (more than a few minutes at a time).  Silicon starts to warp and break down at 90C.  Not that it will abruptly stop at that temp, but that's when the MTBF curve starts to really accelerate.  The card may be qualified to run that hot but that is just in short bursts, not for hours on end.  Even 89 is pushing it in terms of long-term reliability; strive for temps below 80C if you can.

Try to exchange it under warranty, I think you will succeed since you did not modify the card and they should see it as a run-of-the-mill RMA. In theory they could inspect the card and chips and maybe claim it was over-stressed, but that would take a lot of time I doubt they would put that kind of effort in to it since it would be easier just to replace your card.

It may not even be heat related, it may be that you got a card with a bad VRM.

Good luck!

[ThiagoCMC]

Thank you!

[fitty]

92C is really too high for sustained periods (more than a few minutes at a time).  Silicon starts to warp and break down at 90C.  Not that it will abruptly stop at that temp, but that's when the MTBF curve starts to really accelerate.  The card may be qualified to run that hot but that is just in short bursts, not for hours on end.  Even 89 is pushing it in terms of long-term reliability; strive for temps below 80C if you can.

Try to exchange it under warranty, I think you will succeed since you did not modify the card and they should see it as a run-of-the-mill RMA. In theory they could inspect the card and chips and maybe claim it was over-stressed, but that would take a lot of time I doubt they would put that kind of effort in to it since it would be easier just to replace your card.

It may not even be heat related, it may be that you got a card with a bad VRM.

Good luck!

I'm not sure what cards you've used, gpus don't start warping/breaking down at 92C. Sure it's not ideal but 92C won't fry cards.

The board was crap, RMA it, it's not a big deal.

[ThiagoCMC]

I bought a HIS-6990.

http://www.hisdigital.com/un/product2-604.shtml <- Crap?!

I will try to change it for a XFX instead...

[bcpokey]

I think he means your particular card was crap, not the brand. There are bad releases of any component, you were just unlucky enough to get one. But it can happen to anyone, for example when setting up my 3rd mining rig, I put in a Seasonic X750 PSU, one of the highest quality PSUs you can get. Within about 30 minutes I smelled something funny, and about 5 minutes later I heard a loud *POP* and my computer shutdown. PSU was all kinds of dead. Not my fault, not seasonics fault, just a bad psu. RMA'd it and the replacement PSU ran an 800W load for a week before I got my 850W PSU to replace it, and could drop it into a more reasonable load.

[ThiagoCMC]

You guys are awesome!

Thanks!!

[OgNasty]

Ya, don't blame the brand.  I have 2 HIS-6870s and they run fine.  Worst thing that ever happened was when the sticker peeled off one of them.

[sven]

there is a reason most ati cards automatically clock down when reaching 90C i guess

[fitty]

I bought a HIS-6990.

http://www.hisdigital.com/un/product2-604.shtml <- Crap?!

I will try to change it for a XFX instead...

Yeah I meant you just got a faulty card. The failure rate is probably 2-4% for every brand. RMA it and you'll be fine. =)

[kiwiasian]

Love hearing about people who let their cards overheat and then later complain when it stops working.

[Littleshop]

I bought a HIS-6990.

http://www.hisdigital.com/un/product2-604.shtml <- Crap?!

I will try to change it for a XFX instead...

As far as I know all of the 6990 boards are reference now.  Brand almost does not matter.  That will change of course in the future.

[bitebitebite]

 Iv seen 100 degrees when I forgot and left the fan at 35 rpm. Was only hashing at 230 each core when I noticed, but everything was fine when I turned it back up. Couldnt have been more than 30 minutes though. Dont find mining worthwhile now so its pretty safe from here on

[molecular]

there is a reason most ati cards automatically clock down when reaching 90C i guess

untrue. the default fan-controller on my 5970 even controls the fan so the gpu temp is 92°C. Pretty hardcore, aye?

[Kermee]

You weren't the guy who picked up the Newegg Open Box HIS 6990 H699F4G4M within two minutes after it became available @ $521.99 USD when I tried to complete checkout and it was already sold last Friday?

Cheers,
Kermee