• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

GPU Folding Questions

J

Jon855

[H]ard DCOTM January 2008
2FA
Joined
Aug 25, 2005
Messages
12,132
I currently am running the beta version of the CPU clients on my dual core and I'm also folding my 1900xtx. Although I'm not sure if it's working the way it's suppose to work.

Downloaded 6.10 Catalyst - Installed without a hitcch
Installed the GPU Client, configured, ran it - Temps doesn't rise above 42c :confused: me thinks something is wrong here...

I'll give the DirectX an update, the the faggotry WGA is acting gay... Got to make it work...
DirectX Update: I had the lastest version so no need to update the DirectX portion of my system.

So far I've this configured:
Core 1 : idle/100%
Core 2 : idle/100%
GPU : idle/100%

 
Lower both CPU cores to 95% and your temps should go up.
 
Hito Bahadur said:
Lower both CPU cores to 95% and your temps should go up.
Click to expand...
I have killed both of the CPU Instances and ;) my temps hasn't rose one bit as well. I'm polling into this interesting find.
 
Did going to 0% on the cores decrease the time to complete a GPU frame?
 
Hito Bahadur said:
Did going to 0% on the cores decrease the time to complete a GPU frame?
Click to expand...
How would I go about checking the time frame? I'm running the console version as a service. Let me know.
 
Should be an FAHLog in the folder somewhere. It might be named something different, but there is a text file for the CPU versions.
 
Also one important question when configuring on a dual core, do I set it as 3 for machineID ? cause that's what I did.
 
That seems to be recommended although I can't confirm it being necessary.
 
This is a little odd, no matter if I'm running GPU Client or not my temps doesn't seems to reflect the change, checked with ATi Tool and ATi Catalyst 6.10, I don't know what else I'm able to do in order to verify that it's running... Task bar shows it running. I'm baffled. I even also forced my overclocked 3dclocks to run as 2d/3d at all time, with no temp changes as well..

Edit: I'm going to give the 6.5 a try... -_-
 
Suggest running as an App and not a service as some people have been having trouble running it as a service.
 
Hito Bahadur said:
Suggest running as an App and not a service as some people have been having trouble running it as a service.
Click to expand...
That seems to be working now as APP...

Edit: Update, my 3D OC Profile were pushing it to nearly 80c so I revertedd back to stock 2D clocks and the temps doesn't rise above 55c at its peak. Something to think about.

Granted I oc'd from 500Core 2d to 775Core 3d...
 
As per your guide, Current advice is to put the 2 CPU cores at Low setting and put them to 95% load. This seems to makes everything worse... The entire system bogs down. Unacceptable...
 
kk. Just best I have at the moment. If you can get a better setup please let me know.

btw: what is your memory useage? everything bogging down sounds like you are out of system memory.
 
whiny said:
The entire system bogs down. Unacceptable...
Click to expand...

yea, read that a few places

maybe CPU config cores @ lowest @ 95%

GPU @ low @ 100%
 
marty9876 said:
yea, read that a few places

maybe CPU config cores @ lowest @ 95%

GPU @ low @ 100%
Click to expand...
So much tweaking to do... I'll be sure to let you guys all know what works for me superbly.
 
I'll be able to help more tomorrow once I get my system up.
 
Your CPU instances are running as services right? if not, set their priority in task mananger. THey should both be on the lowest setting and the GPU client should be at normal priority.

(Right click on the program under the processes tab and you can see/set the priority.
 
What are your system specs? Dual core I assume.

Best guess, total guess here....

stop all folding

start up GPU, note overall CPU usage

start up client 2 (on 2nd core)

see how system is responding

if ok, add in 1st CPU instance

Note: if you have 3x things on the same CPU priority they will all compete.
 
ehh never set anything to normal I think

also use the configs to set priority, doing in windows never works that well near as I can tell
 
Yeah, just trying to understand what I have seen. Running the CPU clients as services and on idle should put them at the lowest priority.

Running GPU as an APP and (medium?) should put it at about normal priority I think. Should force the two CPU clients onto the 1 core plus taking up leftovers on the second core. That would be the goal, I think, at this point.

FAH really needs to be a single app intelligently using multi-cores/GPU. Not holding my breath though.
 
Based on Marty's post (3 above) You should get the following loadings:

GPU only = 50%
1xCPU(@95%) + GPU = 95%
2xCPU(@95%) + GPU = 100%
 
the run as app not as service boggles me, I think they might be off track there.

I've run FAH CPU version a lot as app, it's running at idle (lowest, what ever you set in config) while in app mode (command box)
 
Hito Bahadur said:
Based on Marty's post (3 above) You should get the following loadings:

GPU only = 50%
1xCPU(@95%) + GPU = 95%
2xCPU(@95%) + GPU = 100%
Click to expand...


short goal, try to get one thing working @ a time... :)
 
Hito Bahadur said:
Based on Marty's post (3 above) You should get the following loadings:
2xCPU(@95%) + GPU = 100%
Click to expand...

Quite close - I'm getting 98-99% I guess this tells us that the GPU isn't using the entire other free 5% of the CPUs.
 
Agreed (with Marty). Just tested that myself and came up with same thing.
 
Jon855 said:
Quite close - I'm getting 98-99% I guess this tells us that the GPU isn't using the entire other free 5% of the CPUs.
Click to expand...

Are you still crawling? How much memory useage?
 
Hito Bahadur said:
Are you still crawling? How much memory useage?
Click to expand...
Am not crawling, I've started a new thread stating what works, CPU at idles/95%, GPU at low100%

memory usage is at 699MB out of 2Gig all of the instances are running big packets.
 
Have you been using 6.10b the whole time? or did you switch from 6.5?
 
marty9876 said:
how many frames have you gotten done at most?
Click to expand...
How would I check this? GPU isn't the same as the CPU where I'll know how many steps has been completed out of... For the GPU client, it only says 1 completed, 2 completed and so on forth... Apparently it's taking me about 5-7 mins per notch. To this date, I've completed only pansy 3, 4's coming any seconds now... I'm running the GPU as an APP right now and its save intervals are at 30mins.
 
I dunno, log do look different. Hard since I don't have one of these.

Try watching less porn?
 
marty9876 said:
I dunno, log do look different. Hard since I don't have one of these.

Try watching less porn?
Click to expand...
here's your porn

[01:33:39] Preparing to commence simulation
[01:33:39] - Looking at optimizations...
[01:33:39] - Files status OK
[01:33:39] - Expanded 83053 -> 443705 (decompressed 534.2 percent)
[01:33:39]
[01:33:39] Project: 2711 (Run 0, Clone 379, Gen 0)
[01:33:39]
[01:33:39] Assembly optimizations on if available.
[01:33:39] Entering M.D.
[01:33:45] Will resume from checkpoint file
[01:33:52] Completed 0
[01:33:53] Starting GUI Server
[01:33:53] Resuming from checkpoint
[01:33:53] Verified work/wudata_05.log
[01:33:53] Verified work/wudata_05.edr
[01:33:53] Verified work/wudata_05.trr
[01:33:53] Verified work/wudata_05.xtc
[01:40:16] Completed 1
[01:46:38] Completed 2
[01:53:02] Completed 3
[01:59:25] Completed 4
 
My understanding is that 1,2,3,4....are the equivalent of frames and supposedly there are 100 of them. 7min 12sec should equate to what they are predicting for a X1900XT at stock speeds (to the best of my knowledge). If you are close to this number then things are going well.

I'm not certain how steady each frame is (so don't know how many frames would need to be completed at different settings to see how they are affecting folding speeds).
 
Those numbers look really good if you are running both CPUs also. Is there a significant slowdown in CPU fold times (time to complete frames)?
 
Hito Bahadur said:
My understanding is that 1,2,3,4....are the equivalent of frames and supposedly there are 100 of them. 7min 12sec should equate to what they are predicting for a X1900XT at stock speeds (to the best of my knowledge). If you are close to this number then things are going well.

I'm not certain how steady each frame is (so don't know how many frames would need to be completed at different settings to see how they are affecting folding speeds).
Click to expand...
Alrighty... I just want the services to work -_- So I can just leave it in the background running day lomg without me looking at the text porn that Marty's sending me.
 
Hito Bahadur said:
Those numbers look really good if you are running both CPUs also. Is there a significant slowdown in CPU fold times (time to complete frames)?
Click to expand...
Both CPUs are running, they each are running at 95% so I'm not too sure how much slowdown that'll show in CPU Folding, I'll keep an eye out on that.
 
Jon855 said:
Both CPUs are running, they each are running at 95% so I'm not too sure how much slowdown that'll show in CPU Folding, I'll keep an eye out on that.
Click to expand...


They are set at 95% but I'm sure one is getting minimal time (is being superceded by the GPU). Let it run for a bit and look at the difference in time to complete frames.
 
You must log in or register to reply here.
Back
Top