|i'm havving problems with the boinc (Boinc Wrapper) results, |
i took the top 3 results from the 100-result sample file and try to reproduce them, but i can't.
the first one (4.385617) give me only a 4.383674, and i think that is what the most get.
ps. i was talking about the Linac900Ext6tc2
[Edited by [DPC]white_panther at 2010-12-25 19:18:04]
|Yes, I've seen some "noise" at the 0.001-0.002 muon% level coming most likely from calculation errors on machines that are e.g. overclocked, hot or not entirely stable. Fortunately the genuine design advances are usually bigger than these errors (CPU errors will typically effect only one particle, not the whole beam). If a result with a very wrong score appears, I'll remove it from the database myself.|
|Looking at the distribution of scores for that optimisation here, you'll see a bunch of marginally-higher scores (off by a single particle probably) but then a big block of 4.383674, which is your repeatable score.|