Thanks Lonewolf. I took your suggestion and am running a GPUGRID WU now. Thank You Sir.

Edit. Note that it says "Running(0.54CPUs+1.00 NVIDIA GPUs). Does that mean that it is using about half of one of my I7Cores and all of the processors in that NVIDIA GTX 580? (hopefully) It jumped the wattage used on that machine from about 160Watts to 335Watts.

snip
Has been for quite some time... several years:

1. Long as you're signed in
2. Long as you've got the direct link to the result log
3. Long as the task is on the RS pages.

Meaning last week the link worked 6-8 hours after validation and in normal operation 1-2 days after validation [see knreed post]. which then also lets the invalid / error result logs disappear for the same task. Now see about half of the last 3 days of validation listed,

So, if support is needed with backup, best to copy /paste whole or part of the result log, certainly the header and ending section where the fail messages are. Been away a day, and you but most importantly support and techs could well miss to see what was there.

--//--

@RT Did you free up one of your CPU cores for the GPU to use? Set BOINC to use 87.5% of your cores (on a quad with HT) and that will leave 7 to crunch and 1 for the GPU. Will let the GPU work much faster since it's not competing with WCG WUs for CPU resources.

agree with Ingleside. I got a few of the first batch so I have turned gpu crunching back off. everything here worked fine.
I have turned the gpu crunching off because I do not wish to crunch the hcc cpu wu's.

It's hard to make sweeping statements regarding CPU configurations due to the variety of CPU's. There isn't a general recommendation for all CPU's, and then there is the GPU variation to consider; there is some difference between a GTS450 and an HD7970.
If you have a dual core, or a quad without HT then it would probably be counter-productive to free a CPU core now. If however you have an 8 thread (or more) system, then the picture changes somewhat. Firstly you could only lose a maximum of 12.5% of your CPU. As operating systems typically use ~1 to 3% when idle that's 11.5 to 9.5%. As other CPU tasks usually benefit from faster access to resources (the CPU is not completely saturated), the loss is usually <8%, and often <3%. On some projects performance running 3 tasks rather than 4 is even more productive. So it's down to what you are crunching. Anyone looking to crunch Beta GPU tasks here could set WCG resource share to a high number (say 10000), leave a CPU thread/core free and crunch other GPU projects (PEOM, MW, Einstein, GPUGrid...). That way they can use their GPU now and when there are GPU tasks here they have a good chance of getting some.
I think I read that one project here suffers from over-saturation of the CPU; task failures are more common when all CPU cores are used. This is not uncommon Boinc-wide. Then there is the other benefits of using less than a full CPU thread complement; the OS/CPU can move threads to keep the CPU cool/less hot. Often one core runs hotter than the others, so the CPU/OS can use the others more. Responsiveness for systems in use is higher, and noise might be slightly less. I find that power is about the same with 7/8threads in use.