Topic: is there a minimum power limit on cards? (Read 514 times)

Boy do I got news for you.  You have the same GPU's as me, MOBO and PSU as me.  I'm telling you right now you can drop the power to them bad boys to 65% and still pump a wicked 23.5-23.75 ETH hash PLUS 180mh/s hash on SC.  Ok, the SC hash is nothing to write home about but the ETH hash is great value.  I have 6x of them babies and my rig runs on 500 watts.  This means you can connect 9 of those beauties to your 1000w PSU.  Return the 850watt PSU.

Those cards punch well above their weight on ETH hash coins but fall a bit short on others (I think its because of the 3gb).

For the sake of all the newbies, numbers for core and ram in the above message are given in MHz, do not try to overclock the memory by 100% (at least indoors)!!!

8 PNY GTX 1060 3 GB    MPN:   VCGGTX10603XGPB-OC-BB  XLR8 Edition
ASUS PRIME Z270-A LGA 1151
Windows 10
EVGA SuperNOVA 1000 G3 powering five cards and the risers, motherboard and SSD
EVGA 850 BQ just powering three of the cards

At this point, I have reloaded the latest Nvidia drivers multiple times. I reinstalled Windows 10. I've updated the bios and checked settings -- I have another rigged with the same exact specs except the cards are not the XLR8 VERSION. i'VE REINSTALLED msi AFTERBURNER.
-I've run only four cards on the PSU 1000 just fine. Changed the riser and card that was disabled -- of course that new riser could be bad also.
-I've run the new card in riser, and three of the cards work fine on the 1000, and 4 other cards on a different PSU. Still got triangle error in device manager on one card.

What motherboard and PSU(s) are you using? Other than a few dedicated mining motherboards most won't run with more then 3-4 GPUs maybe 5.... some can do more then that but they are often a pain in the arse to deal with. At this point we don't deal with motherboards that are not specific dedicated mining mobos, the hassle and headache is not worth it otherwise.

I still can't get that eight card running.

1000 PSU - hooked to risers and 5 cards
850 PSU - hooked to 3 cards


If I try to run all eight cards, I get an error triangle in device manager on one of the cards.

If I run just 4 cards on the 1000 PSU, it runs fine. So, I changed the card and riser out that I had disabled. Nope. I get the same error trying to run 5 cards.

So, I tried running 4 cards and all of the risers on the 1000 PSU. and a completely different PSU 1050, for the other 4 cards. I get that same triangle error again.

It seems to run seven cards okay. However, I was having Internet problems. I am mining zcash. If the Internet goes out, the miner won't restart.it just keeps saying trying to restart trying to restart trying to restart.

Also, my settings in msi afterburner completely disappear now when Windows restarts. I'll try uninstalling and reinstalling later maybe.

70% should be okay

set clocks

to

tdp 70%
core +10
ram + 100
fan 75%

see how that works

[/quote]

I use these parameters, except have the fan on auto. It looked like it was running at about 57%. Fired up five of the cards yesterday at 10:30 AM and ran stable until I went to bed at 10 PM. However, sometime in the night it stopped again. I am mining zcash and using the EWBF CUDA Zcash Miner. In the mining apps window it was continuously trying to restart. I close the window, and tried to restart the minor but it would only flash for a split second and disappear. When I restarted the rig, I pulled up MSI afterburner and all my settings were gone again. So after resetting, I have fired up the five cards again with the  fans set at 75%.

Is it okay for the cards to run the fans at 75% for a long period of time?

Not sure where to go from here. I have tried reinstalling the miner software as well as a refresh copy of Windows 10. Could it still be a hardware issue, even though it ran for so long? still haven't tried another PSU on it.
[/quote]

75% fans are fine

I always set manual fans as they tend to stay put.

dropping settings overnight  could mean :
a bad card
a bad riser
clocks too  high   70 % tdp  +10 core + 100 ram is not too high
do the manual fan setting for 75 over night

 if it fails again

run one  card less   then you are running now see if that works

if you have 1 bad card or a bad riser  on a big rig  say a 6 to 10 card rig  it will take time  to find the problem.

.
[/quote]

70% should be okay

set clocks

to

tdp 70%
core +10
ram + 100
fan 75%

see how that works

[/quote]

I use these parameters, except have the fan on auto. It looked like it was running at about 57%. Fired up five of the cards yesterday at 10:30 AM and ran stable until I went to bed at 10 PM. However, sometime in the night it stopped again. I am mining zcash and using the EWBF CUDA Zcash Miner. In the mining apps window it was continuously trying to restart. I close the window, and tried to restart the minor but it would only flash for a split second and disappear. When I restarted the rig, I pulled up MSI afterburner and all my settings were gone again. So after resetting, I have fired up the five cards again with the  fans set at 75%.

Is it okay for the cards to run the fans at 75% for a long period of time?

Not sure where to go from here. I have tried reinstalling the miner software as well as a refresh copy of Windows 10. Could it still be a hardware issue, even though it ran for so long? still haven't tried another PSU on it.

70% should be okay

set clocks

to

tdp 70%
core +10
ram + 100
fan 75%

see how that works

this is a general question. I am trying to troubleshoot my rig. Should there be any problems with my  cards running if I take them all down to 70% power limit?  Or do certain cards need more power to run correctly? I have GTX 1060 3GB cards.

I figured if I try running them at a lower  power limit that they would be more apt to be stable. Is that right or wrong? I've been working a week trying to get this rig to go. It just keeps restarting or stalling.

Tomorrow I try another PSU on it I guess.