Stephen Brooks 2004-12-19 11:46:47 | So it suddenly jumped massively after a long slow climb. It did this before actually when it went to 0.3%. Is this manual intervention or just a sudden lucky hit by someone's client program? |
Nexus 2004-12-20 05:00:24 | Just wondering the same thing. |
Alex 2004-12-20 06:22:22 | Maybe it's because of the DPC Megaflush? |
[DPC]TeamNWW - Huub 2004-12-20 07:14:41 | I commented on that on our forum yesterday.. Could be coincidence ofcourse but maybe there were some promising strains among the flushed results? |
Stephan Hermens[RKN] 2004-12-20 08:21:37 | Hi there. I noticed, that the first results with a positive percentage were offsprings of the best ChicaneLinacB designs. The parameters of those designs were pretty similar, with random parameters for the linac cells 61 - 90. They still had "linaccells=000" in their genome. So at the end of November I took the best ChicaneLinacB90 design and tweaked the parameters for the last 30 linac cells to match the patterns of the first 60 cells. This design immediatly jumped over the 1% hurdle. I tried 4 other similar designs with comparable results and let my PC run muon without automatic upload. (Only part-time, so not many results were generated) The best design on my PC was 1.107818% when I finally uploaded last Friday (18:42 UTC). Beginning on Saturday the PCs from other participants uploaded designs that were further improved. |
Stephen Brooks 2004-12-21 12:01:47 | This also explains the early jump to 0.3% from negative a month ago - probably when someone put a ChicaneLinacB result inf ro test and got the first 60 cells right. |
DPC_Meloentje 2004-12-21 12:58:17 | That's excactly what I did. I took my best ChicaneLinacB result and placed it as Test in queue.txt and got a .25% result. It also must have been a lucky shot because I wasn't able to repeat this with other ChicaneLinacB results. |