 |
| K`Tetch 2005-09-14 14:05:39 | Its been quiet here the last few days, so i thought i'd post this. I was cleaning up my 'tset' directory (where i test manual designs for a week or so) and looking at my viewresults for the past week or so. i did a screengrab of the phaserotC_bigS1 results.dat historgaph. See if you an spot when the samplefiles were downloaded. http://home.alltel.net/andynoz/samplefiles.png |
| [TA]z 2005-09-14 15:32:01 | are you secretly touting your .2+ % yields?  |
| K`Tetch 2005-09-14 20:16:24 | 0.2 - thats like 'oldskool' style  |
| K`Tetch 2005-09-22 12:27:05 |
Ok, well, its been a week since. Maybe some more explanation might help About a month ago, i sat down, with a manual design from the then top of the samplefiles data, it was about -0.15 or so. I watched it run all the way through, spotted some bottlenecks on the solonoid end, and modified accordingly. did this until i got my first positive result (somewhere about 0.11% yield) and then did two more sims, grafting in my best result from solonoidsto15cm and solonoidsonly, to see how they would fare,a nd give some seperate comparison, and then I left it to the client to improve. I spoke to stephen at this time about it, and my worries about submitting which would corrupt the growth (like the fast jumps in phaserotC did, I'm sure, and i was responsible for the second jump). As you can see at the start of this thread, about 2 weeks ago, I started having the client use the samplefiles, in case there was any improvement with the crossover. I've still been worried about contaminating the results and growth, so far i've returned all my results over the past few days that are less than the current max. however, growth appears to ahve 'stagnated' so i think the time is right to release. My computer is slow after all (look at the benchmark thread to see how fast this dual P3-550 is). Anyway, its now grown in a week almost, so I'm upping them now. Thats basically the point of thismessage, a how and why of the jump for phaserotC_bigS1. as of this moment, my high is 0.492345% - y'all will be faster building it up than me.  oh, for the design junkies, the genome difference (as determined by the genome distance page) shows an RMS difference of: RMS difference = 484.886200879 (max possible 999) between my best and [ta]z's best. Whilst the difference between the top and bottom of the samplefiles file is currently around 66. Big design difference basically. K`Tetch |
