Pages:
Author

Topic: Running MSI Radeon HD 7990's - page 2. (Read 8281 times)

full member
Activity: 167
Merit: 100
December 15, 2013, 11:45:05 AM
#7
I have 17X7990 in total. The core/memory voltage and frequency vary because every GPU is different. For some GPU, I can run 0.99v, with 950/1000 MHz, temperature is 70C. For others, the temperature is 85C, with 0.95V, 870/900 MHz. It depends on the position of the cards in a PC case.

I use powered PCIE risers in open case, I put extra fans to extract heat out.

The original thermal past is extremely thick. I tried to apply thin layer of thermal paste, the improvement is minimal, just 2-5 degree, as the heat sink cannot touch the GPU evenly, even after I adjusted the force with the 4 screws. That could be the reason why manufacturer apply a thick layer of thermal paste.

Thanks for your reply, I very much appreciate it. You've got a lot of hashpower there! Probably about 20Mh/s - I'm just bumbling around at 3.5MH/s.

I might look into reapplying the thermal paste when I have time and the coin craze calms down a bit. Even 2-5 degrees are worth it, particularly if I get 5 degrees less on the hottest GPU. I've also tested the limits a bit on my cards - one card is great and seems to go up to 1000/1000 on 1.02V, but the other cannot cope as well and crashes above 960/1000. I do like to keep my cards cool <72C, so I'm sticking with lower settings. It's all great fun.
hero member
Activity: 896
Merit: 1000
December 15, 2013, 07:35:13 AM
#6
I have 17X7990 in total. The core/memory voltage and frequency vary because every GPU is different. For some GPU, I can run 0.99v, with 950/1000 MHz, temperature is 70C. For others, the temperature is 85C, with 0.95V, 870/900 MHz. It depends on the position of the cards in a PC case.

I use powered PCIE risers in open case, I put extra fans to extract heat out.

The original thermal past is extremely thick. I tried to apply thin layer of thermal paste, the improvement is minimal, just 2-5 degree, as the heat sink cannot touch the GPU evenly, even after I adjusted the force with the 4 screws. That could be the reason why manufacturer apply a thick layer of thermal paste.
full member
Activity: 167
Merit: 100
December 14, 2013, 09:41:55 PM
#5
I am running 4XMSI7990 without problem on Gigabyte GA-Z68AP-D3 ATX Intel Z68 Express DDR3.

I use two OCZ 1250W Gold power supply, which is too high for 4X 7990.

If I run 3X 7990, one 1250W PSU is OK as I undervolt, the power consumption is less than 400W each card.

My setting: 1.0V VDD, core/mem 950/1000MHz. When it runs at 1.2V, the power consumption is much higher than 400W, maybe 550W each card.

The temperature varies due to the quality of thermal paste application on the GPU.

Great to hear someone successfully running these beauties in a 4x setup! Your 1V 950/1000 config sounds very similar to mine. One GPU always crashes at 1V 950MhZ for me, though, so I have to run at a lower 920MHz. I can absolutely confirm the power draw is smaller than 400W, I'm running 2x7990 on a Akasa Venom Power 1000W Gold PSU without any problems. Regarding temps, the GPUs run at roughly 57,60,63,72C, but I've got to admit my current setup is very hacky: built inside an opened case, but with 6x fans mounted around the cards.

I'm really curious how you keep your cards cool? Do you have the normal open miner approach with powered PCIe risers? Do you do anything else, e.g. remove some of the shields around the card?
What temps are you getting? If you re-applied thermal paste, how much did it improve your temps?
sr. member
Activity: 434
Merit: 250
December 03, 2013, 10:02:17 PM
#4
Running more but smaller PSU's is cheaper.

I think the standard setup is 750W PSU for Mobo with 1 GPU and then a 350W PSU for every GPU you add after that.

Keeping a spare 750W PSU and a spare 350W PSU is cheaper than keeping one spare 1100W PSU.
hero member
Activity: 896
Merit: 1000
December 03, 2013, 08:36:24 AM
#3
I am running 4XMSI7990 without problem on Gigabyte GA-Z68AP-D3 ATX Intel Z68 Express DDR3.

I use two OCZ 1250W Gold power supply, which is too high for 4X 7990.

If I run 3X 7990, one 1250W PSU is OK as I undervolt, the power consumption is less than 400W each card.

My setting: 1.0V VDD, core/mem 950/1000MHz. When it runs at 1.2V, the power consumption is much higher than 400W, maybe 550W each card.

The temperature varies due to the quality of thermal paste application on the GPU.
newbie
Activity: 8
Merit: 0
December 01, 2013, 06:19:03 AM
#2
This is probably not much help to you but I've just started mining with a MSi 7990 and at first I thought I would try and get away with my current 700W PSU but no chance, as soon as it got anywhere near full mining speed it would shutdown. I've just installed a 850W and all is good although I'm having serious heat issues at the moment.

I can't run the card anywhere near full speed as I can't keep it cool. Been trying various intensity's in CGminer but it seems to be either half speed or full speed no matter what flags I use, would be nice If I could get it running at around 2/3 of it's max :/
member
Activity: 81
Merit: 10
Litecoinminingparts.com
December 01, 2013, 05:50:29 AM
#1
Hi guys,

Need a bit of advice as I can't really find a lot of information on these online.

What is the maximum number of MSI Radeon HD 7990's I can use on one set up using the 6-rig guide here? http://www.coinminingrigs.com/how-to-build-a-6-gpu-mining-rig/

I have read that using 2 750w PSUs is suffitiant for running 2 7990's. I have not seen anyone using more than 2 7990's and an extra lower powered GPU. For example, the Rocky Mountain miner are using this same setup, 2 7990 and one smaller one so it seems this is the safest way.

Anyone know what MB and PSU they are using?

Thanks,
Pages:
Jump to: