Separate names with a comma.
Discussion in 'Distributed Computing' started by pututu, Mar 22, 2019.
Eric (the admin) has GTX 1050 and his pphr is about 28,740. So a 1050 will produce 689,760 PPD. A 1080Ti will be about 4.5 times faster just by considering the CUDA cores, so roughly 3M PPD assuming it can scale linearly.
Spoiler: GPU cuda 30 run time
If they can build in Linux, i'm sure it'll be ported to x86.
Well certainly never running the CPU app again, lol
Pay no attention to the above posts.
No GPU WU's it's a early "April Fool" joke
I'll work to clear "ALL" these jokes up so y'all can keep crunching the CPU WU's..
Now if I could just get on board with Linux...
It's not that hard
If I can do it anyone can. (after all I've been told I'm not a smart donkey)
Oh, I can run it easily enough. It isn’t difficult, really. My problem is the lack of hardware monitoring programs for Linux. Give me GPU-Z, CPU-Z, AIDA64, and HWInfo64 ported to Linux and I’d run it. Until then, it’s a non-starter for me.
I forget who on our team wrote some of those. I don't think a port is anytime soon tho.
I don't use them (don't remember ever using them either)
The points have been adjusted, it's now 800 points per work unit.
FYI, from the Numberfields front page.
Points have been adjusted again, I do not know what it means for the CPU side, but it looks like PPD for GPU has gone up.