Hi,

my rig configuration is as follow
AMD Phenom X4 9950
4 GB Corsair Dominator
XFX GeForce 9800 GX2

The standard speeds for the 9800GX2 are 600MHz/1000MHz/1500MHz (GPU/Memory/Shader). I am using the latest Nvidia driver 190.38.

I overclocked my card to 726MHz/1193MHz/1815MHz and I put the fan on 100%.

My average CUDA finish time before the overclock was about 15~20 mins. But after the overclock, I am down to 20 Secs per each CUDA unit!!!!

Is this normal? or am I doing something wrong!?

---

anyone offer any suggestions?

---

I have no knowledge because of GPU OC.

I bought 4 from manufacturer OCed GTX260-216.

I would recommend to look to some manufacturer homepages, how high they OC their GPUs.
For example - maybe Zotac or EVGA.
In the time (before ~ 1/2 year) I looked around for the best OC - EVGA had the highest OC for the GTX260-216 of the core/shader/RAM.

But, this from manufacturer OCed GPUs are specially sorted chips.
It's possible that you can't reach this OC with 'normal' stock GPUs.

Also you can look at specially OC sites/forum in the internet.


You have your PCs hidden, so I can't look to the results.
But I guess - if you crunch now the CUDA tasks in only few seconds, then your OC is too high and the GPU make only errors.

You could unhide your PCs in your SETI@home preferences.
Should SETI@home show your computers on its web site? yes


You let run stock_ or opt._CUDA_app with VLARkill?

---

Your overclock is too high and they are having errors.

Expect 11-16 min with a high OC.

Also Memory overclock doesn't help reduce times.

Bob

---

Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957