• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** Official AMD 7990 Owners Thread ***

May need some help, hoping other 7990 owners could input.

The last couple of weeks the card doesn't like mining. I have always under-volted and mined at a sensible intensity as firstly I use the system in the day, secondly I don't want to damage the card.

The last couple of days the monitor is the indicator the display has crashed, when I look at the keyboard LCD the CPU/RAM usage confirms when they do not move. I have to reset the system via the button. One time it has done this and the OS was still functioning as you could see CPU usage just no display.

I have cleaned out the fans and heatsink this morning, some dust but not caked in it. Fans all spin.

When I ran core at 700, voltage 1.2v it mined for a bit then system crashes (no display). There are no minidumps leading me to believe it's a power/GPU issue. Lastly sometimes I can smell something being dumped out by the fans could be burning type smell.

I have yet to test with some gaming until I get some time later on today. Played crisis 3 for 10 mins at stock settings and seemed ok (not great test I know).

Any ideas welcome. Mining wasn't my log term project with this was just doing it to pay off the hardware. Instability did happen a while back with CGminer reporting the API crashing and restarted, lowering settings resolved this. Maybe that is typical if you set your voltage too low?

TIA
 
Core at 700 with 1.2v. Is the stock core not 950?

Maybe 1.2v at 700 is too much. At stock voltage or using LtMatts non throttle BIOS running 100% @ 1000Mhz some can get their voltage down to less than 1.1v.

I'm confused buy the low core value or you mean hash rate? I'm not a miner by the way so not used mining stuff or pretend to understand it.

I'd try mining at stock for a hour or so whilst you are present as the burning smell is an issue. What temps have your GPU's been running at during this mining?

Best
 
Woodsta888 said:
Core at 700 with 1.2v. Is the stock core not 950?

Maybe 1.2v at 700 is too much. At stock voltage or using LtMatts non throttle BIOS running 100% @ 1000Mhz some can get their voltage down to less than 1.1v.

I'm confused buy the low core value or you mean hash rate? I'm not a miner by the way so not used mining stuff or pretend to understand it.

I'd try mining at stock for a hour or so whilst you are present as the burning smell is an issue. What temps have your GPU's been running at during this mining?

Best
Click to expand...

Hey pal. The default on this card is 1000 core 1500 memory. For mining I have only ever run it at 900/1450 with voltage of 1.050 - 1.1

I did not get round to flashing Matt's BIOS as I was wondering if this voided any warranty?

I had to turn the intensity down of one core a while ago as it ran too hot. I set the cutoff levels inside CGminer to turn off the core if it ever exceeded 85c which works well.

So lately the hot core ran Intensity 12, with the better core on 13. This worked up until Monday when I started to notice the lockups. I have read that AB does not apply the undervolt properly if it rounds up, but I am used to manually setting the profile, checking with GPU-Z then leaving it running.

I just had a 30 minute session of BF4 and it was fantastic. I just dont want any ball-ache of RMA especially if they can turn round and point the finger ala "you have updated to a different BIOS", "Removed and re-applied thermal paste" if you know what I mean.
 
I dont think you can void the warranty. You could always flash the BIOS back copying the stock BIOS back on.

I guess you have seen http://forums.overclockers.co.uk/showthread.php?t=18558655 explaining what the stock BIOS does.

You say your card is 1000Mhz but is that not the boost speed? and it throttles between 950-1000? You got the Devil 13?

Seems you know more about it all with all those parameter changes etc and you have had it working. As for AB trawl this thrad and the BIOS throttle thread for AB settings.

You turned the heating up in PC room on the radiator TRV? Surprised they get that hot with lowered clock speeds and voltage. You have a good case? Leave the side off and either dont heat the room or leave the room door open as the card will heat it up. Leave a small window open. Does sound like it has gotten hot. HAs your card sagged? These cards are heavy and they sag - effectively bending the card. And the 7990's come with a backplate, this also bends! resulting in the heatsinks not having the contact they did have raising temps. Either need to lay the case on its side if cooling allows or prop up the card or tie it from above. You may if case allows and you don't mind modding your case get another fan in the case side giving direct air to the card.

They are very heavy cards and I was horrified to see mine has bent and I'm not a miner but a regular gamer.

What are trhe asics of your card as if they are marginally different this can indicate which GPU will run hotter.

Other than that mate you've stretched my limited knowledge. Hopefully one of the other miners of recent purchases may be able to help ya.
 
What power pack have you got/how old?
If its gaming fine you have to wonder about the power supply,mining puts loads of strain on cooling,power and the gpu.... a oldish/cheap power pack might be your issue?!
 
Woodsta888 said:
I dont think you can void the warranty. You could always flash the BIOS back copying the stock BIOS back on.

I guess you have seen http://forums.overclockers.co.uk/showthread.php?t=18558655 explaining what the stock BIOS does.

You say your card is 1000Mhz but is that not the boost speed? and it throttles between 950-1000? You got the Devil 13?

Seems you know more about it all with all those parameter changes etc and you have had it working. As for AB trawl this thrad and the BIOS throttle thread for AB settings.

You turned the heating up in PC room on the radiator TRV? Surprised they get that hot with lowered clock speeds and voltage. You have a good case? Leave the side off and either dont heat the room or leave the room door open as the card will heat it up. Leave a small window open. Does sound like it has gotten hot. HAs your card sagged? These cards are heavy and they sag - effectively bending the card. And the 7990's come with a backplate, this also bends! resulting in the heatsinks not having the contact they did have raising temps. Either need to lay the case on its side if cooling allows or prop up the card or tie it from above. You may if case allows and you don't mind modding your case get another fan in the case side giving direct air to the card.

They are very heavy cards and I was horrified to see mine has bent and I'm not a miner but a regular gamer.

What are trhe asics of your card as if they are marginally different this can indicate which GPU will run hotter.

Other than that mate you've stretched my limited knowledge. Hopefully one of the other miners of recent purchases may be able to help ya.
Click to expand...

Hi woodster it has gone pear shaped now so I will be brief (explanation below).

I dont think this is the mining settings, I have played with different cards and settings since start of 2013 when LTC was easy to mine and less than $2. What I mean by that is I have backed off aggressive so it shouldn't even break sweat equivalent of running 70% I would say lately.

ASIC was 62 and 64% on each core. I posted on this or similar forum of my stilt I have been using to keep the card dead straight.

The API crashing a few weeks back leading to me backing off and now the current situation lead me to believe it's a PSU or GPU issue.

Bazippy said:
What power pack have you got/how old?
If its gaming fine you have to wonder about the power supply,mining puts loads of strain on cooling,power and the gpu.... a oldish/cheap power pack might be your issue?!
Click to expand...

Hi Baz. I got with the card an XFX pro black edition 750w which is made by Seasonic. I have been running off the wall with a Power Monitor and never run the rig above 500w. When AB did not apply the voltages properly I would catch it with the monitor but even then it would not have exceeded 600w.

--Update--

OK so I went to a job this afternoon and returned to find the machine on a windows recovered from a serious error screen. The machine was left on idle only running a trade script which is even more worrying.

Reboots managed to system restore then since it doesn't manage to fully get back into windows as it keeps on rebooting. I am in the process of testing RAM, HDD etc from USB typing this on the laptop.

I am going to swap a spare GPU in soon and then the 650w PSU before I have to raise an RMA. Looks like the issues I noticed a couple of weeks back were early warning signs of something not right.

Anyway if you can think of anything it may be or of help I appreciate it.
 
It was your stilt I copied. Thanks! Apologies for not making connection and getting you to suck eggs.

Not good news fella. Guess you'll have to go through the process of elimination like you posted above. Really tests your diagnostics skills when something like this happens before you make the RMA decision. You are lucky to have your spares.

Good luck
 
Bazippy said:
Considering doing this for summer,never done any watercooling but mining is off at the moment once summer comes!
What did you get etc?
Click to expand...

EK full cover blocks (directly from them), two EK 360 radiators and the EK pump/res 4.0 combo.

Will be all mounted into a water cooling test bench chassis as the cards are used for mining at the moment and I want to keep them as cool as possible.
 
Woodsta888 said:
It was your stilt I copied. Thanks! Apologies for not making connection and getting you to suck eggs.

Not good news fella. Guess you'll have to go through the process of elimination like you posted above. Really tests your diagnostics skills when something like this happens before you make the RMA decision. You are lucky to have your spares.

Good luck
Click to expand...

Thanks.

Before I swapped the card out I decided to re-arrange the modular power cables around. The two 8-pin PCI-e supplies were off one cable, so now it has a dedicated feed from the PSU to each port.

I have been mining with it back to stock settings then started to tweak again. Current run has been going now for 1 day 9 hours.

I think I must have been undervolting the cores too much which caused the instability. I am hoping the arrangement change from the PSU is what that burning smell might have been from but it seems to be crunching away happily even though one core starts to hit the 90's. Only other possibility was a poor connection, and I have removed the additional HDD's/pci card to do this testing.
 
Th0nt said:
Buzz does not have his trust setup, was going to ask him how the 7990 is fairing since he tamed his temperatures?
Click to expand...

Sorry man, not been very active recently. My temps are as good as I will get them without water cooling and very much under control. I have not had a single crash outside of Battlefield 4 and due to a monitor upgrade it's running at 1440p now.

I haven't actually done anything since I put that piece of wood there, been thinking about maybe using a zip tie to hold it in place and make it look a bit neater inside, but to be honest I rarely look inside anyway unless I am adding/removing parts or just giving it a cleanout.

I have also sorted my trust now, so you can get in touch if you need me.
 
FirebarUK said:
Lately I've been finding that when Windows turns off my monitors the fans on my 7990 ramp up to 100%. It continues until I wake the displays. Anyone else found this before?
Click to expand...

Sunds like a bug that occurs when the monitor goes into standby mode. Try disabling it and just turning the monitor off instead. Does it fix it? I had a similar issue way back when i was using a 7950 crossfire setup. At the time i was overclocking via BCLK and stopping that fixed it for me.
 
LtMatt said:
Sunds like a bug that occurs when the monitor goes into standby mode. Try disabling it and just turning the monitor off instead. Does it fix it? I had a similar issue way back when i was using a 7950 crossfire setup. At the time i was overclocking via BCLK and stopping that fixed it for me.
Click to expand...

I had a similar problem when my monitor went to sleep, I had a blank screen when I woke the monitor up. Disabling monitor standby mode worked for me, never had a problem since. hopefully new drivers may fix this.
 
Need some advice..iv never watercooled before but iv made so far about 2k in a few months mining but the temps get very high and im tired of having to leave my windows open all the time just to try and cool it down..What exactly do i need to watercool just the GPU and what would be a good one to go for...
 
New issue!

I have now managed to install 13.12 buuuuuuuuuuut

x116 BSOD when trying to play BF4/Bioshock but only when I have my CPU overclock set.

No overclock and I can play fine. Any pointers?
 
Unfortunately 14.1 crashes for me in the same areas. I think it's an unstable overclock crashing the display driver. But I need to find the right setting.
 
You must log in or register to reply here.
Back
Top Bottom