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.
So I've managed to get all my GPUs to allocate to all 4 cores, what I am curious about has anyone found a certain core to be better or is setting them to all 4 the best results?
[H]ard|DCer of the Month - February 2007/January 2
Joined
Sep 24, 2002
Messages
5,028
I found I get less lag if I have the GPU on core 7 and the VM's on 0-6.
But I have to isolate the 4x VM's on cores 0-6 so that Windoze and the GpU client share core 7.
If I isolate the GPU client on core 7 and have the VM's and Windoze on cores 0-6, the computer lags like fun.
Running everything on all 8 cores still lags but not as bad as isolating the GPU on core 7.
Running XP64 Pro and useing SMP Seesaw Pro to set affinities.
When I tie my GPU to all 4 cores my dual smp's still play nice with it using the spare cycles. It really is all in setting the GPU to higher priority and the SMPs to idle.
I'll have to try core 0 and 2 setting on a machine and see what that does for the GPU.
The system that i've been testing on nets the following: (Q6600 @3.2Ghz
No Affinity for anything IE everything set to 0-3
GPU 4557
SMP1 855
SMP2 931
Affinity changer on and GPU set to 0-3
GPU 3165
SMP1 968
SMP2 931
Afinity Changer on and GPU set to 0,2
GPU 3165
SMP1 968
SMP2 931
Seems like this paticular system is happier with affinity changer off. I am going to try a single SMP later today to see if a single SMP with AC plus the various GPU configs = the an increase in PPD.
[H]ard|DCer of the Month - February 2007/January 2
Joined
Sep 24, 2002
Messages
5,028
Which version of Windoze are you useing, XP or Vista ??
Which protiens are you crunching, a fast or slow one ??
Whats the SMP's PpD without the GPU client ??
If your running XP what happens when you lock the two SMP clients to only 3 CPU cores.
If I lock two of my VM's to only 3 CPU cores of my Q6600 @3.1 Ghz under XP, they only loose ~20% of their speed, 2,000 PpD -> 1,650 PpD while crunching p2605.
So I'd get 3,300 PpD plus the GPU client out of this box.
5214 on the GPU, Windows XP 32bit, SMPs is 2665's, hits about 3k when using only SMPs and AC. The problem is there is little way besides doing VMs for me to mandate the GPU stay on say core 2 while the SMPs take the other 3. Unless you know of some method.
Hmm might have to try VM's again not against it, they just seemed to not net enough gain for the hassel before, but now the ability to limit to CPU's would be beneficial.
[H]ard|DCer of the Month - February 2007/January 2
Joined
Sep 24, 2002
Messages
5,028
If it a pure folding box then try the isolate function of SMP Seesaw Pro.
Once you isolate a process to a core/cores with SMP seesaw then nothing else will use that core/cores unless you manually set affinity to let it.
So try isolating the GPU client & move the GPU core on one CPU core.
Then Windoze and the SMP clients should only use the other three CPU cores.
I'm running on 32 bit XP as well.
I had been using NotFred VM SMP clients for several weeks before the GPU client came out. When the GPU client arrived, I toyed with several configs trying to tweak for max production, and seeing Tigerbittens' adviice on using SMPSeesaw with the VM's I followed it. The GPU client locks to core 3 on a Quad and core 1 on a Dual so I just used SMPSeesaw to "move" both VM's to 0-2 and left other running processes (including Folding@Home.exe) to wander on 0-3.
This seems to give me the best output (considering it's my main rig and surfing slows the GPU client down a little) at similar numbers as TB. Currently the VM SMP's are pulling 3200 with a 2605 and a 2665. It's a little higher with both running 2605 of course. Plus the GPU.
When I started using the VMs I ditched AC for SMPSeesaw, as AC seemed to be geared more to native Windows SMP clients. But SMPSeesaw requires you to set it when you start things running, so I don't see any way to automate it for reboots/power outages. this affinity manager may take care of the "automation" part, but I haven't tried it yet.
Q6600 system at 3ghz.
4gb RAM
Vista 32 bit
1 GPU2 Client
2 SMP Clients
Affinity Changer
So Monday evening I installed some Windows Updates needing a reboot. I figured I turn AC off to see how my performance did. This affected the 2 SMP clients pretty hard. It seemed to make no difference to the GPU client. For the past couple of weeks the GPU client has pretty much been rock solid at right around 4400ppd. Overall the machine had been averaging 7400PPD. With AC turned off it dropped to about 6600PPD. Both of the SMP clients took a significant point drop. Last night I turned AC back on. This morning I check and fahspy is now reporting 7400PPD again.
Not sure why my config uses it different but I guess we all need to experiment little to see where you will get the best result. Any questions on my config let me know.