New folder i7 3820

T

tcclaviger

n00b
Joined
Jan 7, 2012
Messages
55
Currently have my rig setup as follows:

i7 3820 - 4.9 ghz
16GB DDR 3, 2044 mhz, 9-11-10-28-1t
2x 6870 1060core/1125ram

Have F@H setup to run SMP -7, and 2 GPU units, then adjusted the "Percent CPU Usage" to 87%, so that 1 core can still run both GPU projects.

My first WU for the CPU I have gotten is a PRCG 7610 for SMP and 11293 for both GPUs.

SMP
Estimated TPF SMP: 7min58sec
Estimated TPF GPU: 3min07sec (x2 of course)

Does this seem right for my hardware? I feel it seems lower than it should be, but I have nothing to compare to except random forum posts I can find on forums. The CPU is stable, the GPUs are stable and it demolishes any and every benchmark I have thrown at it.

9.79 Cinebench

This was before the bump to 4.9ghz, obviously.
9-11-10-28-1Tat2022.jpg
 
Last edited:
Looks right to me.

Be careful running -smp 7. Some of the newer WU freak out with that number of cores. Just keep an eye out for any failed WU. If you get some, switch to SMP 6 or 8
 
The 7610 are some of the longer slog smp WUs around. Looks good.
 
I will have to pull up my stats. Honestly with radeon 6800's I would say dont bother folding on them and just go full hog on the CPU. You will definitely improve PPD/W and possibly even improve overall PPD as well.
 
extide said:
I will have to pull up my stats. Honestly with radeon 6800's I would say dont bother folding on them and just go full hog on the CPU. You will definitely improve PPD/W and possibly even improve overall PPD as well.
Click to expand...

This is the way I would go. But if you have cheap/free power, go for it!
 
Thanks guys. I will try both, with and without the GPUs crunching each way for a couple days see what I get overall.

Power cost isn't an issue for me thankfully.
 
if by PRCG 7610 you mean project P7610 your TPF seems off.
My i5-2500K @ 4.5GHz (Nobby) has a TPF of around 4m45s and a Q6600 @ 3GHz (mustrum) is around 8m30s, both under win7



Uploaded with ImageShack.us
 
You do not need to cpu adjust if you are running smp -7, cause that leaves 1 core free.
 
Thanks for the heads up on the TPF times. After disabling the GPU slots and setting 8/100% the TPF times are down to 3min05sec :p

Now to find a project with very little CPU overhead that is GPU only.
 
tcclaviger said:
Thanks for the heads up on the TPF times. After disabling the GPU slots and setting 8/100% the TPF times are down to 3min05sec :p

Now to find a project with very little CPU overhead that is GPU only.
Click to expand...

I think one of the DC commandos should fill you in where your ATI cards can make some waves ;)
 
I was doing RC5-72, aka dnetc but decided that this was far more worth of time than a simple key brute force effort.
 
not too sure bout the science, but for creating some waves you with the ATi's you could always check out the [H] team on Collatz Conjecture (which is running under BOINC), but I think the guys in the DC Vault thread here on the forums know more about that one.
I am running that on my measly 5770 and it doesn't seem to impact the CPU that much.
 
If you have problems with smp 7 try smp 6 and a uniproc slot - no point in wasting spare resources:D
 
I have a 7611 in my database, getting the following on a 4Ghz i7 920


Min. Time / Frame : 00:04:00 - 20,594.5 PPD
Avg. Time / Frame : 00:05:16 - 13,631.3 PPD
 
You must log in or register to reply here.
Back
Top