Folding@Home: THGC Needs You -Team 40051

Page 41 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
WHY am I only getting the points for 1 card?

Just checked updates and its 1 unit every time?? :fou:

What have I missed?

Does anyone know if I have to use the -forcegpu nvidia_fermi switch for the 460's.

What does it do anyway.
 


Hold on, need to bring out the books for dual gpu-ing.
icon_e_ugeek.gif


I explain what you need to set up in a minute.
 
http://www.stanford.edu/group/pandegroup/cgi-bin/edit/English/WinGPUGuide#ntoc4

Ok, for your questions, this link should provide all the info you need.

Now, from what i see, you may need to add the "-forcegpu nvidia_fermi" switch although i would think you would only need to add that if you have g(t/ts/tx) 2xx cards and a fermi card.

Now what i think your situation is, (thats if you got 2 gpu clients going) then you need to add "-gpu 0" on one client and "-gpu 1" on the other client to run on different gpu's.

If you only got one client going, only one gpu is going to run. You need the same number of clients to run on the same number of gpu's your going to run. This means if you 8 gpu's (4 dual gpu cards) you'll need 8 clients to run all gpu's.
 
I used the guide to set it up and did all the necessary -gpu0 etc last night.

I also cleared the work and queue.dat folder so I don't duplicate.

Both GPU's are folding 99% usage.

I have 2x shortcuts for each card.

I can start one of either cards or both together.

They are both registered under hfm with individual start end finish times etc.

The only other thing is under hfm part of the bar is orange and it says this client is working on the same unit as another client, for some reason it seems like they are folding the same unit which would explain the points.

Under hfm it also reports each unit as being completed by each card with the points for each card but having checked the eoc page it only had 1298pts per update and 1 unit.... :fou: :fou:

What time is it where you are warmon?
 


I am waiting. :)





Just a quick check: You do have two different FAH folders (one for each card), correct? Also make sure the "Start in" location of the shortcut point to the right folder.
 
^Ha :lol:

Yep, 2 folders, 2 shortcuts, start in pointing to gpu and gpu2 storage folders.

I've just deleted all the work an queue folders again and restarted, the orange bar that said they were folding the same unit has gone so fingers crossed.

Only thing is my username is now orange and it says the user name isn't the same as the configured user name!!??

So you don't think I need -forcegpu nvidia_fermi flag, what about -advmethods??

[:bohleyk:1]
 


[:randomizer]

Thinking, thinking, thinking, thinkingthinking,

You did set the client ID to different numbers on the gpu's? Im guessing so but just making sure.

Beyond that. I cant think of why they would be folding the same thing. I haven't messed with dual gpu before so im out of ideas of whats going on. [:isamuelson:8]

 
Yes, annoying that I folded all night and only got the points for 1 card. [:fisshy:2]

Now that I have deleted and restarted where hfm has "Project(Run,Clone,Gen)" the codes are different.

I'm hoping this has sorted it but I'll find out when the updates are done.

Any ideas on the username problem or the -advmethods flag?
 


Seems plenty of people do, my mates running a 4870 and a 4850 on his, for what its worth. :lol:
 

I didn't say it wasn't possible, just that I've not had much success myself with getting it running with a kind of stability.
 


When I said "for what its worth" I meant points wise, just incase you thought anything else.

Anyway good news, deleting ALL the work and queue.dat folders has worked.

Maybe it just got confused with all the old data being present.

Now we Foldin. Thanks All [:bohleyk:1]
 


So, do you feel the power yet? Muahahah!!!! 😀
 



:heink:

7k base points or with bonus? Whats the WU number?
 


Nope thats what im looking for. 😉

6040, very nice WU's and the largest ones for SMP2. There the closest you'll get to Bidadv WU's without working on them.

Only reason why i dont like them is the fact that they take 3x as long a the normal WU's i like. (1hour per frame vs 20 mins per frame.)



 
Screenshot2.png

[:badge:3]

And yet all my math teachers have said there could never be a "-"0 [:mousemonkey:5]

I was just waiting for the stats to come up and when i refreshed the page. I saw that negative 0.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom