 |
| Stephen Brooks 2007-11-20 17:29:13 | Currently PhaseRotEby5 is on about 0.8 and DecayRotB on 1.2. If it keeps on at the current rate we might be looking at it overtaking in about 3 more days, though there is certainly the possibility of it slowing down. |
| Stephen Brooks 2007-11-23 11:54:01 | 0.977 vs 1.195 right now. Eby5 has slowed down a little. I'm not sure yet whether my changes to Eb6 have made it faster or not, there's the possibility of a 3rd E lattice, probably replacing DecayRotB when it is overtaken. |
| Stephen Brooks 2007-11-24 00:38:15 | DecayRotB is retired, it wasn't really going anywhere and it was taking up a large portion of the computing power due to its comparatively longer simulation runs. But we can just assume it would have stayed at 1.195% or thereabouts for a couple more weeks, judging by recent behaviour. |
| Stephen Brooks 2007-11-24 19:34:29 | Eby5 is on 1.165% but does not appear to be exploring the longer channels (larger values of decaycells and phaserotcells) and neither is Eb6. The increase in yield and overtake of DecayRotB will be welcome, but I might try to add another lattice that might make extending the channel length easier. |
| RGtx 2007-11-24 20:39:48 | Yesterday, I fiddled with the number of Decaycells, setting the parameters of s15, s20, s25... to those of s10, finding that each increase gave an increase in final muon% at the expense of a massive increase in Mpts/run. My current highest result of 1.134064% (Decaycells=874) took 1502.2Mpt. One problem, I found when experimenting with such high values for the number of Decaycells, is that the graphical client seems to stall at the startup splash, making it difficult to judge my progress, and to determine which parameters to include/omit as neccessary. |
| Stephen Brooks 2007-11-24 21:16:18 | Hmm the client hanging is odd. If you e-mail me a genome that causes this I'll test it at work on Monday. What is interesting is the current best in the file I saw still had low values <100 for both "cells" parameters. I want the client to be able to increase these on its own but so far it doesn't seem willing to do so! So I have to keep thinking of different ways of rephrasing the problem in the lattices. |
| RGtx 2007-11-24 21:25:06 | "Initiating quarantine of score 1.45993". Should I separate all such results, and run the client offline for a few days to see how they progress, without polluting your gene pool? |
| Stephen Brooks 2007-11-24 22:43:49 | Nah send them in. If I'd got the lattices working right the client would have done this itself, at least this way I know there was some manual intervention. |
| RGtx 2007-11-24 23:56:25 | Graphics problem resolved - It would hang on every simulation - corrupt display.dat file, replaced and all OK - unfortunately now overwritten, so cannot see what had happened. |
