[01:01] MSG: Quit: Trillian (http://www.ceruleanstudios.com [07:55] Join: fiveop joined #corewars [09:45] Join: Mizcu joined #corewars [13:23] Join: michal joined #corewars [13:26] Join: Core29 joined #corewars [14:44] MSG: [16:38] MSG: Quit: michal [16:53] Join: Core29_ joined #corewars [16:54] Join: sf joined #corewars [16:54] Hi [16:56] MSG: Ping timeout: 252 seconds [17:02] Join: michal joined #corewars [17:02] MSG: Remote host closed the connection [17:14] Does this count as the origin of the term troll? [17:14] http://www.ociw.edu/~birk/COREWAR/88/HILL/fastvamp31.red [17:28] hrm.. [17:28] Hi Bvowk [17:28] what'd you do sf?! I can't get on nano anymore.. [17:28] you're nasty... its just not fair :) [17:29] It would be interesting if your warriors had a line like ;strategy spl mov mov mov mov [17:29] Sorry. [17:29] heh [17:29] you wouldn't be that interested... [17:29] they'd all be much the same. [17:29] I'm close to getting on the hill with a warrior with 48% ties [17:29] wow.. [17:29] thats impressive [17:30] I just need to get it up to 30% wins [17:30] I'm going through each for, trying to get one on the hill [17:30] (until I get bored) [17:32] each form? [17:32] As called by Jens [17:33] spl mov mov mov mov, etc [17:34] I'm having problem getting a mov mov spl mov mov on the hill [17:34] Join: Mizcu joined #corewars [17:35] I think its called by me actually :) I gave jens the list [17:35] although he's done much better than I :( [17:35] :-) [17:35] Hi Mizcu [17:35] hi sf [17:35] Where is he now on the hill though? ;-) [17:36] true.. [17:36] now you're using my list to beat me too! *sob* [17:36] I don't have your list [17:36] I'm just going through some possibilities [17:38] oh.. you can have the list if you like :) [17:39] Ok, thanks [17:39] you could also do a little creative greppage on the 152k list to get your own copy [17:39] :-) [17:40] Join: Roy joined #corewars [17:40] hi all [17:40] Hi Roy [17:41] Congrats on Grendel, looking great on koth [17:41] Thanks [17:41] Join: michal joined #corewars [17:44] Hi Michal [17:45] hi [17:46] hi [17:52] 50 29.3 24.4 46.3 Bombus Mocsaryi S.Fernandes 134.3 1 [17:52] :-) [17:53] They are getting better (not! :)) [17:53] Polaris is pretty good, 22nd.. [17:53] Thanks [17:53] Handwritten? [17:53] Or evolved? [17:53] Now I'm trying to get something with high ties on the hill [17:54] All handwritten except Terrestris (which was from a script but not exactly evolved) [17:55] Mine are all evolved...except the Scanner [17:55] That's dropped down a bit now :-( [17:55] Quite a bit yeah, but well...my evolved warriors are gaining ground again [17:56] Especialy Go on! that one was hanging on the bottom, and now going strong again [17:56] I have oneshots and scanners on there [17:57] Plus some variations with spl mov djn in various combinations [17:57] Yeah, those do well [17:58] spl mov mov spl mov might do well, maybe I'll try that [17:58] Yes I was thinking about that, with the last mov and imp [17:58] yup [17:59] O have tried 4 x mov with 1 x spl with the spl in every position apart from the last. They all make it onto the hill [17:59] Hehe, nice [18:00] This one with the spl in the middle gets most ties [18:02] MSG: Read error: Connection reset by peer [18:08] Join: jkw joined #corewars [18:08] hey jkw [18:09] Hi jkw [18:09] hey bvowk. my comp is powered on once more. [18:09] hey sf. [18:09] MSG: Remote host closed the connection [18:10] jkw: could you point the address for the faq in the rec.games.corewar koth statuses to the new faq please? :-) [18:11] jkw: http://freebiology.org/wiki/REC.GAMES.COREWAR.FAQ [18:21] Join: Roy joined #corewars [18:38] Roy: that's not the name of a real species! [18:38] it is..! [18:38] a new one :) [18:39] http://www.nhm.ac.uk/research-curation/projects/bombus/alphabet.html [18:39] sigh :-) [18:41] It was buggy anyway [18:42] Join: michal joined #corewars [18:44] See how much difference a real name make? ;-) [18:44] Yeah :) [18:50] Join: johnkw joined #corewars [18:53] * sf waves [18:53] Part: Roy left #corewars [18:53] MSG: Quit: Bersirc 2.2: All the original sexiness of Bersirc, open to the world. [ http://www.bersirc.org/ - Open Source I [18:53] MSG: Ping timeout: 252 seconds [19:19] Join: Judodudo joined #corewars [19:20] hello [19:22] hi [19:27] Part: Judodudo left #corewars [20:05] Join: Fluffy joined #corewars [20:06] :) [20:08] Roy: Interesting name for your nano warrior :) [20:12] michal: Are you there? [20:17] sf, I'll look in to that. oh sf is gone. [20:17] Join: sascha joined #corewars [20:18] Hi sascha [20:18] hi 2 all [20:24] Fluffy, I am here [20:24] michal: Do you have any special hints for optimizing a MARS? [20:25] make lots of special cases [20:25] already done [20:25] take a look at martin ankerl's mods to exhaust [20:26] are they in the current version of exhaust? [20:26] no [20:26] Then this might be a problem, because his pages about cw are no longer online [20:26] exhaust-ma in google [20:26] ok [20:27] One more question ... [20:27] or exmars, the core is the same I think [20:27] o.k, just here for a question...must go on....maximum tired from this day...cu [20:27] How to implement a process queue? At the moment I'm using a simple array with indicies to the core [20:27] * sascha waves [20:27] Part: sascha left #corewars [20:27] Do you know any better way? [20:27] * Fluffy waves [20:27] (too late) [20:27] martin stores pointers to the core [20:28] exmars is pretty much exhaust with the order of evaluation redone.. [20:28] as do I after learning this [20:28] so for most of the instructions, they can be evaluated much quicker.. [20:29] pointers to the core ... nice idea ... thanks :) [20:29] well [20:29] he uses pointers all the time instead of indices [20:29] thats true too.. [20:29] one trick I came up with [20:29] for the process queue [20:29] they're pretty much functionally equiv tho.. [20:29] I can drop in either one and my code runs without any changes.. [20:30] just faster with exmars [20:30] is to just allocate enough memory for maxprocesses*2+somemore [20:31] and get rid of folding the pointer (end+1 -> begin) [20:31] this works good for middle sized core [20:31] for a tiny core it might be a bit worse [20:31] if the whole process queue would fit in l1 cache [20:31] but usually it doesnt :) [20:32] Thanks for the ideas. Plenty of things left to optimize :) [20:33] fluffy: you really want to make sure your handling of the mov is first rate.. [20:33] its about 99.99999% of the evaluations.. [20:33] (ok.. maybe thats overstating a bit) [20:33] Unfortunately I had to "delete" the complete c-code and rewrite it. I couldn't find the bug in my original code [20:34] But at least everything works now ... only at 50 % of the speed of pmars :( [20:34] and only iCWS '88 [20:34] ouch! [20:35] I wanted to write clean and easy code in order not to make another bug again, I cannot find [20:35] Now I can start to optimize it. [20:36] bvowk: Let's wait a little bit. Everything faster than pmars is good enough for me [20:38] bvowk: Of course I've forgotten to mention, that it has a nice Python interface [20:38] nice :) [20:38] so its C w/ a python harness or is it mostly python? [20:39] something in between [20:39] interesting. [20:39] Everything, that needs raw cpu power is implemented in C [20:39] want the code? [20:40] yeah [20:40] you know my email right? :) [20:50] hey Jens, since you're such a python guy.. did you want to look at id? [20:50] my *not* evolver? [20:50] I'd be interested in suggestions.. [20:51] yes, I'd like to have a look at it [20:51] please use the gmail address [20:51] I don't want to start my "big" computer [20:51] k.. [20:52] it uses mysql for the backend.. [20:53] likely more beef than it really need.. [20:55] Before I forget it: Everything should work in PyCorwar, but there are still some FIXME's left to be done [20:55] Multiwarrior-fights ('88) should work, but positioning the warriors isn't properly implemented yet [20:59] SF: www.math.ualberta.ca/~bvowk/files/form-list.txt [21:00] its sorted most common to least common, which roughly translates to most effective->least effective, with a little uncertainty thrown in by the evolving.. so its most effective/most resiliant -> least/least. [21:03] looks interesting [21:11] MSG: Quit: michal [21:18] Now I'm trying to understand the MARS in exhaust-ma .... quite interesting :) [21:29] I can no longer think, need some sleep. (Not, that I can really think, if I don't need sleep ;-) [21:29] * Fluffy waves [21:30] Part: Fluffy left #corewars [21:38] MSG: Ping timeout: 252 seconds [21:43] Join: Metcalf joined #corewars [21:43] Hi [21:45] Martin Ankerl said I can put exmars on my site until his is back up [21:45] Unfortunately, my computer has died [21:53] Luckily, I have quite a few important bits and pieces on my USB drive. [21:53] I've been wondering for a week [21:56] I looked inside and the hand-drive doesn't have a connector like any I've seen before :-( [22:01] MSG: Quit: humhum [22:03] Does anyone know the origins of the terms Paper / Scissors / Stone? [22:03] I know about the warriors from ICWS 1990, I just wondered if there are any earlier references [22:07] * Metcalf waves [22:07] MSG: Quit: mov.i #1,1 [22:13] Join: elkauka joined #corewars [22:13] Part: elkauka left #corewars [23:19] MSG: