Post By Thrasher
Justed added 8 cores.
Just added 8 2.2's for you (Xeon E5's). All eight cores are 100%, no other tasks or runtimes on machine, 24GB RAM.
Anyone know what to expect from it?
how many PPD?, how long per unit? etc...
never Mind... just got the estimate... was blank for the longest time... (about 17hours/unit, we'll see how that pans out tomorrow), 17,000PPD
that's seems very low, I think I was much higher on another project with a dual GPU rig,
anyone chime in and give me your thoughts here, I'd rather run the GPU rig if it will be much faster than this unit.
Zero GPU power in my rack, but lots of procs and memory.
you could always assign the cores into groups instead of running a single unit on several cores. also the estimate will fluctuate until it gets a few units down as the Client first needs to run a few units for efficiency estimates.
but for comparison im running an nvidia 650 ti and averaging `14k -15k a day so.......... YMMV
Im sure poking around the official F@H forums you may find optimization setting for cpu's in general but in the end the design of the GPU cores is more suited for crunching numbers, as this is primarily what they are designed for anyways.
Last edited by Thrasher; 11-14-2013 at 04:00 PM.