You are viewing a single comment's thread from:
RE: Gridcoin mining experiment overview
Wow, 41 views and 100 votes as of now.
Similar fluctuations in RAC value are visible in other projects, too. BOINC credit system attempts to reward work measured in FLOPS. The fundamental problem is that FLOPS is not easy to determine or measure in advance for a given WorkUnit.
RAC in Amicable Numbers project also shows similar randomness. In case of Amicable Numbers project the cause is unpredictability of WU parallelisation level read more here.
Have you found what is causing RAC fluctuations in VGTU? I assume some WUs are computed faster than expected and others slower, thus different pts/s yield. If that is true, why the same hardware is processing some VGTU WUs more efficiently than others?
I've seen your articles and read them ;-)
For VGTU I'm almost 100% sure host normalization is involved. I've written about it here. I expect this to happen with other projects as well, but I haven't any info on those. Some will, some won't I guess.