*** Official AMD 7990 Owners Thread ***

[BuzzLloydyear]

I'd check the seating/thermal paste of the heatsink. When I took the cooler off my Gigabyte card to put on the water block one of the memory chips still had the plastic backing on the back of the thermal pad! You may improve things by re-seating everything....

To be honest, I would rather try to RMA the card. I have not had any luck with it since I purchased it. When it performs it's great, but between keeping temps in control and all the many issues I have had with it. I would rather see if I can replace it instead.

I just did Heaven bench on stock clocks, stock fan profiles, ULPS off constant voltage off (which ran heaven, but then caused my pc to crash shortly after). By the end of the bench, gpu1 was hitting 93 degree's @ 77% fan speed, gpu2 was sitting comfortably at 75 degree's @ 44% fan speed.

That is with constant voltage off, which I absolutely can not go without as it causes the pc to become unstable and will blue screen or lock up when playing games.

There has to be something wrong with it somewhere, it just seems to be too troublesome. If it's not one problem, it's another. I haven't had a single period where everything has been working great. It was a bargain when I bought it at £480 2 months ago, but I have had nothing but problems with it since I bought it. The only redeeming factor is even when it's not working, it still outperforms my 6950, so it was still a step up. For the money I paid though, I shouldn't have to be in hardware management mode as often as I am.

 

[gpuerrilla]

djay does your 620 have two 8-pin connectors?

I think my 650 may only have two 6-pin cables. Man that's cheesed me off as I don't think my new PSU is going to arrive till monday.

Inspected more closely - they have 6 pin blocks with a 2 pin attachment so looks like I will have a wee bash.

 

[djay]

Inspected more closely - they have 6 pin blocks with a 2 pin attachment so looks like I will have a wee bash.

Hi, yes mine has the 2x6 pin connectors then each one has the smaller 2 pin attachments hanging loose. Just allows you to use either 6pin or 8pin whenever necessary.

I still can't get MSI AB to show the FPS overlay on screen when running anything. I have follwed LtMatts guide but nothing seems to work.

 

[gpuerrilla]

Those that are mining LTC. What settings are your config files on?

Oh and coming from two 7770's do I need to re-install the drivers or flush any changes?

 

[Bazippy]

Im using guiminer for litecoin

thread concurrency 14336
gpu threads 2
worksize 256
intensity 13
getting average of 1.3 mhash

Coreclock at 950 and mem at 1450
dropped voltage to 1150 to try tonight then try lower tomorrow(still getting 1.3 at the moment)

 

[ToxicTBag]

I guess I am back on the amd bandwagon 2 x 7990's inbound .... no intention of playing Crysis

 

[Bazippy]

struggling with heat from one let alone two

 

[daveip]

I'd check the seating/thermal paste of the heatsink. When I took the cooler off my Gigabyte card to put on the water block one of the memory chips still had the plastic backing on the back of the thermal pad! You may improve things by re-seating everything....

If you're willing to do this, I think this is the way to get your GPU temps down + more consistent.

I'm on water now, but with a 5% ASIC difference between my core 1 + 2, temps between both cores are never out by more than 2C when I have them both running.

And on a side note ... litecoin mining ... hmmmmmmmmmmm

 

[Bazippy]

Temps are reasonable so don't think i need to reseat anything,just higher then my old 6850 cf setup

 

[gpuerrilla]

Im using guiminer for litecoin

thread concurrency 14336
gpu threads 2
worksize 256
intensity 13
getting average of 1.3 mhash

Coreclock at 950 and mem at 1450
dropped voltage to 1150 to try tonight then try lower tomorrow(still getting 1.3 at the moment)

When I set to intensity 13 my wattage goes over the 650 supply in there at the moment so after 10 mins I backed off to 12. I suppose I can undervolt using AB or it it locked till flash to Matt's BIOS?

 

[Bazippy]

I installed the beta version of msi and i have voltage control,no bios flash (dont want to flash because at 1.3 im higher then the 7990 is usually rated so im not messing with it!

 

[CaptainRAVE]

LtMatt you are a star. Running at 1.120v (although this isn't quite stable, so will increase it a little), GPU at stock and all sensors being monitored on both GPUs. Your bios runs like a charm and the instructions were very straightforward. Amazingly my temperatures now rarely cause my fan to go above 30%. Very impressed.

Edit: 1.130v seems stable

 

[Bazippy]

I installed the msi beta and gained voltage control,no new bios

 

[Bazippy]

oops didnt mean to post that twice

 

[daveip]

Im using guiminer for litecoin

thread concurrency 14336
gpu threads 2
worksize 256
intensity 13
getting average of 1.3 mhash

Coreclock at 950 and mem at 1450
dropped voltage to 1150 to try tonight then try lower tomorrow(still getting 1.3 at the moment)

How many shares are you connected to in order to get 1.3mhash?

I'm only managing 1.0mhash using the same settings.

 

[Bazippy]

Dunno,just set those in and it worked

 

[daveip]

Are you quoting the mhash rate from stats on your account on the pool or from GUIMiner?

I'm showing 1.2 on the stats on the pool but GUIMiner's showing 1.0? :S Go figure ...

 

[Bazippy]

yeh ive had that before,gui quoting 1.3 average on pool is 1.3 roughly and getting peaks as high as 1.5

 

[BuzzLloydyear]

If you're willing to do this, I think this is the way to get your GPU temps down + more consistent.

I'm on water now, but with a 5% ASIC difference between my core 1 + 2, temps between both cores are never out by more than 2C when I have them both running.

And on a side note ... litecoin mining ... hmmmmmmmmmmm

I don't doubt that this would help with the temps a bit, but the temps is one of a few issues I face. The other one is I have to force constant voltage for the card to remain stable. Even at stock settings, without that set, I get random blue screen's when not gaming and when gaming my pc will lock up. I can't explain why that is. My only assumption is due to my 1st gpu core being poor, it is not reacting well when the cards voltages drop. By any measure, forcing a constant voltage adds more heat and is not really a fix. It's a compromise, which I shouldn't really need to make considering the cost of the card (even now and I was unfortunate to buy it 8 weeks ago for £100 more).

I will see if I can send mine back. At present I don't believe it's of sufficient quality and the amount of work and effort I have put in is beyond what I would expect already.

It's already been more trouble than it's worth and even though I have a stable performance (if limiting due to temp control and constant max voltage). I shouldn't be making compromises to the level I have, to get it to where it is now.

 

[ikari1]

Shocking.

Really would like to take mine apart and re paste it but it was warranty void stickers on four of the screws. Saying that though my temps are pretty much identical so maybe maybe its already a good un and i should leave well alone.

Mine are exactly the same temp at the same useage. To be fair I can't remember what they were running on air for the short time before H2O..... What I can remember is they would touch 90degrees C at 100% load. Frightening! Now they don't get much above 50degrees C on BF4 running 90%+ @1150 core.