Run Time vs CPU Time - what is going here?

[blckgrffn]

MalariaControl is one of the projects I have a goal to hit 1M points in. I have two X6's pushing towards this. One has ridiculous "run times" versus its "CPU Time" - the other seems to be reasonable:

http://malariacontrol.net/workunit.php?wuid=59823402 <-- 1090t (7,502.99 4,889.86)

http://malariacontrol.net/workunit.php?wuid=59822340 <-- 960t (2,917.81 2,887.56)

Now, here is what they were doing in PrimeGrid:

http://www.primegrid.com/workunit.php?wuid=244949378 <-- 1090t (1,850.01 1,821.26)

http://www.primegrid.com/workunit.php?wuid=244920968 <-- 960t (2,061.97 2,029.65)

In the meantime I have changed one thing - moved from the stock cooler to the CM 212+ on the 1090t. Is it throttling or something? Will Thubans do that?

Later tonight I will reboot the 1090t host and change the "target" temp from 60 to 50 or so to see if that will motivate the cooler... it's running ESXi so no temp monitoring, unfortunately. Not that I could get lm-sensors working on the 960t computer with the same motherboard working anyway...

Thanks for your insight in advance 🙂

 

[GLeeM]

Do you have the GPU running anything that uses quite a bit of CPU? And don't leave a core for the GPU WU?

I get similar results on the thread that shares CPU with the GPU WU.

 

[blckgrffn]

Nope, this baby is just CPU.

Wait.

HAHAHA.

F@H client started when I rebooted it 😛

Good call, turning that baby off 😉

It was pulling an epic ~600 ppd fighting BOINC.

Thank you for asking the right question GLeeM, it was key to me solving the problem 🙂

 

[petrusbroder]

LoL, I was going to propose that when I read the OP: 🙂