 |
| Stephen Brooks 2008-12-16 18:11:40 | This is the first user to submit a positive Linac900Removable3 result, well done. [TN]shauge 2034 123938.4 (17.1%) 0.132647 |
| [TA]Assimilator1 2008-12-18 23:53:47 | Is this a new optimisation? |
| [TA]Assimilator1 2008-12-19 00:27:44 | Lol, I should of read the title properly!  So what's the '3' refer to? |
| tomaz 2008-12-19 08:18:46 | Stephen, graph of "12" doesn't work, it is not clickable, so results can't be retrived. |
| Stephen Brooks 2008-12-20 01:58:48 | Sorry, the script that generates the "clickable" results retrieval grid only runs once per week (I think) because it's quite time-consuming. It works now, but only generated half an hour ago. The number refers to the number of "segments" the linear accelerator is divided into. Earlier we had optimisations where every component was different but that stopped working well once I included the accelerating components, so I simplify slightly by treating them in blocks, but I can choose how "granular" this is. |
| [XS]riptide 2009-01-10 05:53:38 | I see a major jump in both 3 and 12 recently. Stephen and Tomaz ran a couple of quantum (yes pun I know) leap simulations to bring it forward. >2.9% now |
| Stephen Brooks 2009-01-12 14:20:32 | Yes, I seeded with the best from the "1" lattice. Now they will diverge from that point. |
| Stephen Brooks 2009-01-14 15:02:39 | Interesting observation: now that I've relocated the focus of the optimisation to an already high yield, the "12" optimisation is improving faster. One explanation would be that in this territory it benefits from having more parameters to tune. |
