EVGA

V7 and Mixed GPU - Fermi & GPU2

Author
krillind8z
Superclocked Member
  • Total Posts : 147
  • Reward points : 0
  • Joined: 2010/04/09 19:18:36
  • Location: Las Vegas
  • Status: offline
  • Ribbons : 1
2012/11/17 16:15:26 (permalink)
Hopefully this will help a few people out that were having problems with V7 Client and running 8, 9 and 200 series cards with 400, 500 and 600 series cards.
 
System Specs:
Evga P55
i5 750
slot 1 16x 480GTX
slot 2 16x Empty (for breathing room while folding)
slot 3 4x 260GTX
NVIDIA Driver 306.97
 
I'm using V7.2.9. I opened it up and immediately hit pause to stop all of the clients for the time being. Opened up precision to show the GPU usage graph so when I started to test I could see exactly which GPU was being used when starting the GTX260 folding. Right clicked only on the 260 and hit fold.

 
I noticed that GPU1's clock started to climb. GPU1 is my 480GTX. So I clicked paused and clicked on Configure.  Then clicked on Slots and went to the first GPU in the list.

Clicked Edit and opened up the Folding Slot Config. Here is where I had to change the OpenCL-index and Cuda-index. I had to change them both from -1 to 1 (still not sure why exactly since it's in slot 3 you would think 2 but I think this is where all of the confusion comes from)

Once I did that I clicked OK then clicked Save on the Configure screen. I tested it to see if it would fold. It immideately started folding on the GPU2 client and the GTX260 was at full clock and usage hit 99% in precision. (Forgot to take a screen shot of it being the only 1 running). I didn't have to delete anything from the work folders since it had the unit just assigned to the wrong GPU which was just fixed.
 
After that I changed the the GTX480 which is showing as GPU1.
Went back into configure and highlighted Slot 2 and clicked Edit.

Since I already found out that the 260 had the Cuda-index and OpenCL-index of 1, I set this one automatically to index of 2 and added the extra slot option  client-type  beta.

 
OK out of that and clicked Save on the Configure screen.
Right clicked on ID 2 (Slot 2) and clicked fold. Here is a screen shot of them all folding. I paused it to do a few things with screen shots and started them all up and took this one.
 

 
The reason why I have 4 different IDs showing under the Work Queue is because I normally have 2 GTX 480s in there. I paused them, I didn't have a QRB unit at the time so it was perfect, and removed 1 of the GTX 480s since I don't have a large enough power supply in there for all 3 cards. Once the GPU2 unit finishes the other 480 is going back in.
 
 
Hopefully this helps anyone that was frustrated with V7 and their older cards!

Main PC- HAF-X - ASUS Sabertooth x79 - i7 3820 - GTX680     
2nd PC - HAF-XM -  EVGA P55 SLI - i5-750 3.2GHZ - GTX480 SLI
3rd PC - Antec 900 - ASUS P5N32SLI - E6600 - GTX260_216CORE

#1

1 Reply Related Threads

    TheWolf
    CLASSIFIED Member
    • Total Posts : 3841
    • Reward points : 0
    • Joined: 2007/11/14 16:05:23
    • Location: Pascagoula, Ms
    • Status: offline
    • Ribbons : 9
    Re:V7 and Mixed GPU - Fermi & GPU2 2012/11/17 16:34:01 (permalink)
    Very nice lay out & instructions. Thanks for posting..
    I see where my problem was when I tried to set this up now.
    I was thinking like in old v6 where -GPU 0 is used so I was using 0 for 1 of the slots instead of 1 & 2 for slots.

    EVGA Affiliate Code ZHKWRJB9D4 My HeatWare 
     
    #2
    Jump to:
  • Back to Mobile