[02:03] MSG: Ping timeout: 245 seconds [07:59] Join: fiveop joined #corewars [15:43] christ [15:43] somebody needs to go through optimax and give it REAL error messages [15:44] somebody needs to make a version 1.0 of optimax [15:48] 80,69533868The system cannot execute the specified program. [15:48] 80,69533868Returncode : 256 [15:48] that error means [15:48] "you can't have a warrior with a space in the filename" [15:48] i think. [15:49] :) [15:50] 'course i'm still tracking function calls [15:50] now it's in another file [15:51] which of course shows up like shit in notepad [15:51] >:( [15:52] * pak21 suggests not using notepad :-p [15:52] Hi all, by the way... [15:52] i was correct! [15:52] hi [15:53] lame [15:53] i don't want to fix the perl though [15:53] my $params = join ' ', 'optimars', @_, '>output.lst'; [15:53] @_ i guess is all the params, but lord knows how they are divided [15:55] guess it wasn't spaces [15:57] filename too long [15:58] ok there you go linuxfags [15:58] both scenarios being something i cannot fathom [15:58] it seems to be some windows error [15:58] it won't run with those params [15:58] shorter filename fixed it [15:58] * myndzi puzzles [15:58] NOW i get to copy it all over into virtualbox and hope nothing breaks [15:58] OHAY [15:58] i got a bsod [15:58] from accessing the fileshare [15:58] woot. [15:59] you should've knocked wood [15:59] i didn't even run optimax yet [16:00] SUCCESS [16:00] only took me like an hour [16:00] totally not gonna be worth it too ;) [16:01] i just want to see if this idea works [16:01] and now i must get my ass to work too [16:02] note: it also didn't like me changing the name of the directory "work" in the optimax options [16:02] apparently it's hardcoded in some places [17:28] Join: CoreOld joined #corewars [17:36] Join: yoR joined #corewars [18:02] Join: evitable joined #corewars [18:19] Hi yoR :-) [18:20] I'm working on passing you in recmath :-P [18:20] And you're doing a great job :) [18:21] My new algorithm with the Gaussian matrix processing doesn't really work [18:21] There are a lot of primes... [18:22] All of mine for n>8 are generated by hand [18:23] Wow, really? Thats cool [18:30] I've taken the lead :-P [18:31] Who needs computer optimisation! [18:36] taken the lead of what? [18:36] RecMath ;-) [18:36] overall? [18:38] No, just passed Roy [18:38] oh [18:43] hrm. [18:43] Are you planning to enter bvowk? [18:44] possibly. [18:44] work is a little crazy [19:09] MSG: Ping timeout: 245 seconds [19:13] * yoR is ahead again ;-) [19:13] Thanks to JM's pointers (thank you!) [19:21] if you are allowed a set of n numbers and have less then 2n-1 primes you're doing something wrong :) [19:21] MSG: Ping timeout: 245 seconds [19:29] Join: Metcalf joined #corewars [19:31] i didn't quite gather what the thing was with the numbers [19:32] you have to select N (8, say) numbers such that [19:32] every number that can be generated by adding or subtracting one or more of them from each other in all combinations [19:32] produces the most primes? [19:33] (or say, 3: a, b, c, a+b, a+c, b+c, a+b+c, a-b, a-c, b-c, a-b-c, a+b-c, a-b+c) [19:40] myndzi: yes [19:40] who's in the lead now yoR? [19:41] Metcalf: not you [19:41] oh yes you [19:41] check again ;-) [19:42] metcalf can you give me a list of how many primes you have for each n? [19:42] Hehe nice one JM ;-) [19:42] I think a nice optimization strategy would be: [19:43] given n you try to look for as many primes as possible from n-1, giving extra weight to numbers that are already primes of or produce primes of prime pairs [19:43] then add a 2 [19:48] It's better to have one odd number and the remainder even [19:48] yes probably [19:48] :) [19:53] and the values are bounded by 2^32 [19:53] (though perhaps the math isn't) [19:57] wowsah, 2.2 gigs for all dem primes [20:15] thinking about it, this should actually not be all that hard [20:15] i dunno how long it'd take to compute though [20:16] the max number of primes you can get out of any set of N constants is 2^N ya? [20:16] wait, i guess i didn't think about subtraction [20:31] MSG: Quit: humhum [20:34] only about half the combinatinos can ever be prime [20:34] i think [20:34] 'cause if all the constants were prime they'd just add up to a bunch of even numbers [20:34] and if all the constants were even, they'd also add up to a bunch of even numbers [20:35] so what metcalf said with one odd number and the rest even, [20:35] it kinda reduces N by 1, and that many combinations is the max that might be prime(?) [20:35] since you have to 'reserve' one number for making odd numbers with [20:37] an even number of odd numbers doesn't help you because you can't use any of them to make odd numbers from any combination they both are in [20:37] but then 3 odd numbers..? i wonder [20:37] there are less combos that involve 3 of them [20:37] I checked with 3 odd numbers, but there were less possible primes [20:46] yeah, i was just thinking [20:46] possible primes is basically (excepting 2 as a constant), equal to the number of possible odd numbers [20:46] which can be figured out pretty easy [20:47] is 1 odd number better than 5, 7, etc too? [20:48] Yes [20:48] For the number of possible odd combinations [20:49] a plan is forming in my head slowly hehe [20:49] but i can't hold it all at once! [20:53] oh, now i get it [20:53] :D [20:53] i wonder how slow the program will be [21:44] i wonder, can you include a particular number more than once? [21:44] You can, but it's a waste [21:44] because you can't count a prime twice [22:46] MSG: