Yes, i noticed too that i had no comlpete crash since the update.

---

i just had a unit stall on me, and a reboot fixed it.

---

my cuda-z report runnig at

---

I have had CUDA running since my 1st post to this thread, gpu fan set to 75% ( bit noisy) and temps stable at 42 degrees C. So far no snow on the screen and no driver resets by Vista. I have probably just jinxed it by posting this......

---

Bambi

That is actually GPU-Z which works with all cards, not just CUDA flavors.

---

That is actually GPU-Z which works with all cards, not just CUDA flavors.

the link (blue text) is cuda-z report

---

Just got a driver reset, I assume it occured during processing of seti cuda and not gpugrid.

Some observations:

The current gpugrid task "time sliced out" in favor of seti cuda and I processed about 10 or so seti cuda before gpugrid got served back up. This happened automatically and the following image is what I observed shortly after installing 181 drivers on Vista 64


Note that GPUGRID is %60 done and waiting to run and SETI CUDA is running. I checked about 3 - 4 hours later and observed two driver reset popups and confirmed Computation Error here: http://setiathome.berkeley.edu/results.php?hostid=4642025

GPUGRID is back running at 63%, seemingly without error, and I still have about 10 or so SETI CUDA that are waiting from the original 20 or so.

OK, I assume that a project gets what, an hour or so to run?. I would have thought that as soon as one of the SETI CUDA's finished, that GPUGRID would get sliced back in. This worked out pretty good as it did process a bunch of SETI CUDA then switched back to GPUGRID automatically.


EDIT: want to mention that all day yesterday, not a single SETI CUDA was downloaded. Not sure why after about 2-3 GPUGRID that the seti project allowed about 20 or so of their CUDA's to download.

hmmmm ... one between two normal running tasks ....

WU true angle range is : 2.711142
SETI@Home Informational message -9 result_overflow
NOTE: The number of results detected exceeds the storage space allocated.
Flopcounter: 2119552843.983235
Spike count: 0
Pulse count: 0
Triplet count: 31
Gaussian count: 0
called boinc_finish

---

I noticed (as did a user on GPUGrid) that it only rcognizes one card in dual configs. I have 2 cards in a crossfire bord. Had to revert back to 180.48 to get both recognized again.