[00:10] Part: Mannerisk left #corewars [00:10] Join: Mannerisk joined #corewars [00:10] Part: Mannerisk left #corewars [00:18] Join: Mannerisk joined #corewars [00:41] bleh [00:41] this game is so hard >_<;; [01:24] on the contrary, it's one of the easiest games I've played [01:24] it's the setup that's hard :D [01:29] -_-; [01:29] I think [01:29] this is the only game I have played were it is as fun to learn to play as it is to play [01:29] maybe even more so [01:31] Part: Mannerisk left #corewars [01:48] omgg [01:48] I just figured it out :D [03:23] Join: Mannerisk joined #corewars [05:24] ogod [05:24] is there some way to do if statements besides using JMZ and JMN? [06:10] nevermind lol -_-; [06:12] Join: X-Scale joined #corewars [10:07] Join: Neogryzor joined #corewars [10:36] oh shit [10:37] maybe Metcalf has archived version of the page [10:43] http://web.archive.org/web/20071219024634/http://vyznev.net/corewar/guide.html [14:09] cool, more new players :D [14:09] put some stuff on the SAL beginner hill so i don't feel like i'm playing against myself [14:09] :> [14:11] there's one free slot there.. [14:12] with quirin going down, theres gonna be easy scores for a paper (even for an old-style paper) [14:15] i'm not that far into the meta game yet :P [14:15] hi [14:15] hi Neo [14:16] i see you are again on top of the beginner hill [14:16] its not like i ever wasnt [14:16] stoney was getting near to the death-age, so i replaced one that was less defensive [14:17] though it scores even better, which wasnt really my intention [14:17] :D [14:17] laf [14:17] "i can't help it if i'm good! meanies!" :P [14:17] ? [14:18] a free spot... thats a temptation... [14:18] Hint: Crackhammer is weak against big decoys, specialized bombs and backward-clearing oneshots [14:20] and Wumpus finally got an opponent that it can be proud against [14:21] what is Wumpus? [14:21] hehe [14:21] Wumpus = upgraded raidar, raidar was first CLP-style warrior [14:21] it scores well against that! [14:21] (and now Myndzi got headache because he doesnt know what CLP is) [14:22] i read it but i've forgotten [14:22] :) [14:22] CLP = Curious little program, made by Kline [14:23] Basically a Core clear wich detects other incoming clears and copies itself away [14:23] well, i did know what raidar was so i assumed it was something like that [14:24] Kline's CLP got notorious because he included it in a p-spacer, and the p-spacer ended up getting around 90% losses against some opponents [14:24] got stuck in CLP mode or something? [14:25] because CLP had a brain-washing clear and lot of opponents switched to oneshot-strategy when brainwashed [14:25] lossess -> wins, means to write [14:25] hehe [14:30] if JR is reading, Sol could use an imp-spiral for support [14:45] Join: baudchan joined #corewars [14:45] huh, that didn't help at all [14:45] hehe [14:45] what didn't? [14:46] played around with my best warrior and brought it down a notch :P [14:46] ah [14:46] I am still writing mine lol [14:46] myndzi: spl-bomber? [14:47] ? [14:47] oh [14:47] incendiary [14:48] hmm [14:48] what does it look like? :o [14:48] incendiary, or the warrior? [14:49] warrior ;\ [14:49] how do i answer that question? [14:49] he probably wants to keep the source for himself [14:50] ah [14:50] oh, i don't really mind sharing, but i'd be embarassed [14:50] :> [14:50] but my guess is mov/mov/add/jmz/coreclear [14:50] I didn't know if it was even a valid question ;\ [14:50] your guess would come pretty close ;) [14:51] nothing interesting about this one [14:51] i'd rather score well with something interesting [14:54] the one I am working on right now is supposed to scan backwards till it finds a bit that is not zero, and then (assuming it is the bit that will be executed next) places a DAT there >:3 [14:54] amicool? ^__^ [14:54] place more dats! [14:54] (is dat sum scannr) [14:56] really it was just designed to kill a crappy test bot I made, so I am happy with it. [14:56] I mean, for my first piece of code it isn't terrible and does its job most of the time -_-; [15:03] just tell me when you start feeling cocky with your code, and ill tell how it scores against a benchmark [15:03] lolk [15:04] because that is like being given a double-leg dropkick on the balls [15:04] I would suspect so [15:07] like [15:07] I am stumped on this piece of code that isn't working like I want it to ;\ [15:07] Crackhammer scores over 200 against beginner-hill, 150 against weak benchmark, 100 against good benchmark, and maybe 75 against 94'nop (havent tested) [15:08] if you want me to have a look at it, just send it to some pastebin and drop a link here [15:09] ya [15:10] heh all i was trying to do was add a decoy [15:11] but now i have to re-maths the thing [15:11] i tried to leave the pointers behind [15:11] and i can't make it bomb itself properly [15:11] maybe i'll just copy the pointers after all [15:11] probably a better idea anyway, but i was hoping to make it smaller [15:11] i was curious if it'd take #1 with a proper decoy ;p [15:13] MSG: Quit: Leaving [15:15] hm i could just use jmz as a fallthrough [15:15] perhaps that's a better idea [15:24] observe the balmer effect; http://pastebin.ca/raw/1091829 [15:28] like [15:29] subt SUB subt, 1 is equal to SUB 0, 1 which i dont think is what you intended to do [15:29] yeah [15:29] use SUB #1, 1 (or, SUB #anything_but_one, 1) [15:30] kk [15:31] the problem is [15:31] and you dont need that jmp 2 | jmp -3 structure, just jmz -1, subt [15:33] i've got a line [15:33] mov.i @1, >-7 [15:33] and the instruction at -7 isn't being incremented at all [15:33] wtf? [15:34] are you moving something with b-field of zero? [15:34] that causes the increment to be overwritten with a zero [15:35] apparently so [15:35] myndzi: you broke something [15:35] indeed [15:35] well, i knew it was broken [15:35] i just saw the behavior and was confused [15:35] (jeez, sherlock, you sure are smart today) [15:35] but that's just 'cause i forgot indirection again [15:36] lols fail [15:36] i just liked what it drew! [15:36] i'll unbreak it too [15:36] just experimenting [15:36] i mean, #2 is pretty fun and all but hey! [15:37] oic, solution presents itself [15:40] myndzi: you might want to start using benchmark to evaluate the changes you make [15:40] www.koth.org/wilkies/ [15:40] okay, so B on the JMZ command will incriment as it should, but the MOV uses the number of commands behind itself subt is (which is -2) [15:41] (the benchmark-program that comes with wilkies is kinda useless unless you code in command prompt like i do, just use challange/whateveritwas in corewin) [15:42] would my problem be an addressing mode problem? ;\ [15:42] I dunno addressing modes very well yet lol [15:42] in short, yes [15:43] kk [15:44] http://users.evtek.fi/~mikaos/a.red [15:44] (or JMP -3..) [15:45] loop: SUB #5, 1 [15:45] start: jmz.f -1, -7 [15:45] MOV 2, @start [15:45] JMP loop [15:45] heh, almost the same [15:46] (jmp -1, <-1 is not what i meant. jmp -1, <-2 was..) [15:49] i think faster than i write [15:57] MSG: [16:04] also I see a lot of people putting DAT at the beginning of their code [16:04] mine dies when I do that ;_; [16:04] because those people give a different starting location [16:04] (when none are specified, the code is assumed to start from first line [16:04] there are two ways to do that [16:05] first, label the line where you want to start (well, you dont need to, but that helps, a lot) [16:05] oic [16:05] then either add ORG LABEL at start of line [16:05] start of code, i mean [16:05] or END LABEL at the end [16:06] yeah [16:06] for some reason it didn't occure to me that people were doing it that way lol [16:06] -_-; [16:11] MSG: Quit: This computer has gone to sleep [16:20] Join: elkauka joined #corewars [16:20] hi@all [16:20] h [16:27] nice to see all the new players :) [16:32] yes [17:16] grr [17:17] started coding jmz-scanner in case sal-b might need one in future, but mod-4 plus transparency plus scanner equals pain [17:19] started hacking a mars in pure python with focus on minimum instruction execution to save cpu cycles to run on google app engine ... equals pain too [17:27] MSG: [17:28] Join: baud joined #corewars [17:41] grr, this scores horribly (80 wilkies) [17:41] Join: baudchan joined #corewars [17:42] why two clients? [17:43] I keep switching between my desktop and laptop [17:44] and I close my laptop when I am done getting stuff off of it [17:44] right [17:44] indeed [17:44] MSG: Client Quit [17:44] sorry :\ [17:46] 95 wilk, better [17:48] what is a wilk? :o [17:48] short for wilkies, a benchmark from '94 [17:48] usually the first one i recommend for beginners [17:49] icic [17:49] there are wilkies, wilmoo, wilbez, wilfiz, fshnop (version 2 and 3) and wilmiz under construction [17:50] wilkies was named by John Wilkinson that put it together, and the joke got stuck when wilmoo came out [17:51] indeed lol [17:52] im planing Wilmiz to replace Wilmoo and Wilbez; i dont like them because they include p-spacers which corrupt the scores when fighting non-pspacers [17:52] like [17:52] I thought I would make stuff that could defeat pretty basic attack patterns [17:53] and then when I grow more competent I would take whatever was at the top of the hill [17:53] and fight that [18:05] anyone know where that article is that lists operands and modifiers? [18:05] I was reading it a few days ago and seem to have lost it [18:06] http://www.koth.org/info/pmars-redcode-94.txt [18:06] thanks [18:06] (there is actually one newer than that, but that is good enough) [18:10] MSG: Remote host closed the connection [18:13] hmm [18:14] (no, that wouldnt work) [18:14] but im off to play ET, shout if you need me [18:21] Join: cywhale joined #corewars [18:21] hi [18:22] hello [18:22] how may i be of assistance? [18:23] just read about elkauka coding a mars in pure python [18:24] well, then im not needed in here, back to game [18:24] tried this about a year ago, 'phobos' is a basic mars supporting some of pmars' cli options [18:26] but it's one or more bug(s) - dwarf vs. dwarf 10 rounds results in something like 1/9/0 all the time... [18:27] Tried to adapt the ICWS mars.c structure to python, this is working but definately slow, about 22k instructions per second. [18:27] Part: Mannerisk left #corewars [18:28] Maybe elkauka would not mind sharing some ideas about saving cpu cycles ? [18:29] The basic idea behind phobos is developing a corewar ide like corewin and ares using wxPython for linux systems. [18:33] Join: baudchan joined #corewars [18:36] MSG: Client Quit [18:37] Join: baud joined #corewars [18:55] MSG: Remote host closed the connection [19:22] Join: Mannerisk joined #corewars [19:29] Part: cywhale left #corewars [20:21] Join: baudchan joined #corewars [20:21] MSG: Client Quit [21:05] Join: Dwaddle joined #corewars [21:05] hi there [21:11] Join: boaf joined #corewars [21:14] hi [21:14] how may i be of assistance? [21:15] completely clueless ;-) [21:15] well, the links at top should be useful [21:15] I stumbled over the game googling for linux systemcalls [21:16] our main tutorial has been nabbed by domain-squatter, but this still works http://web.archive.org/web/20071219024634/http://vyznev.net/corewar/guide.html [21:16] I'll have a look [21:17] Mizcu: Thank you. I've been looking for that document for hours. [21:18] Also, I'm interested in compiling pMARS (or a more suited simulator, if there is one) for Mac OSX. Is this a possibility? [21:19] it is possible, but requires some changes in the source [21:19] unfortunately i cannot help there, but newsgroups should have enough information to do it [21:19] (just use groups.google.com) [21:19] I see. Thank you. :) [21:54] hey.. python.. [21:54] use pycorewars :) [21:55] if you need a copy (I don't know if fluffy still has it about) you can email me. [21:56] Join: Metcalf joined #corewars [21:56] Hi :-) [21:57] hi [21:58] i though about CLP with mov.i {xxx, yyy | add.f #xxx, -1 inside the loop, but it think thats kinda impossible [21:58] Mizcu: I'll try to contact Ilmari / upload my backup of his page tomorrow [21:58] I guess the impring is broken now :-( [22:00] It doesn't sound too bad [22:01] Have you looked at all of the mini-CLP variations? [22:02] i considered the pair to be too chaotic to be actually implemented [22:02] unless you use really wide a-field [22:02] Mizcu: I've downloaded the pMars simulator and ran a few testwarriors,this is a great game. I going to wrestle myself through the docs and make a simple and probably stupid warrior myself and test it ;-) [22:03] Hi Dwaddle [22:03] With my knowledge, mov 0, 4 dat sounds good enough. ;) [22:03] Hi Metcalf [22:03] Dwaddle: feel free to come here and ask if anything comes up [22:04] Mizcu: I will come here, with a lot of questions [22:05] But first I will read the docs an do a bit testing [22:05] Hi boaf [22:05] Hello Metcalf. :) [22:06] Nice to see new new people this weekend :-) [22:06] It makes me sad that CW has been around for ages and I'm just now learning about it. [22:07] Same here. [22:08] I stumbled over it, looking for something else. Never heard of it before ;( [22:08] MSG: Quit: http://www.mibbit.com ajax IRC Client [22:09] Join: boaf joined #corewars [22:09] :/ Random disconnect. [22:09] its the webchat that does it [22:10] I wish I'd discovered Corewar at least 10 years earlier [22:11] Metcalf: 3 warriors changed in Wilmiz-draft; http://users.evtek.fi/~mikaos/bench.txt [22:18] one warrior that would fit nicely into the "other" category would be OOS, but it scores too well, shame. [22:36] MSG: Ping timeout: 245 seconds [22:43] Join: elkauka joined #corewars [22:48] hmm.. idea? [22:49] though still not what i'd want.. [22:58] @cywhale : first check out fluffys pyCorewars (isnt pure python though, uses c extensions) [22:58] @cywhale : http://corewars.jgutzeit.de/pycorewar/download/index.en.html [22:58] @cywhale : i made a wx frontend ala corewin for it, if find the code i put it up somewhere ;) [22:59] wow, it's getting more and more crowed it here ... did i miss a note in the mainstream media that mentions corewars? :D [23:03] I heard about this game a year ago or so [23:03] just getting around to trying it out [23:07] gee, the corewars conference was 4 years ago ... im getting old :( [23:14] hrm, isnt linked at the frontpage anymore, but a bit of google-foo reveals the magic: http://www.corewar.info/meeting/ [23:14] hrm. [23:15] how do I use the benchmarks? [23:15] Mannerisk: what simulator do you use? [23:15] CoreWin [23:15] though I think I have pMars around here someplace [23:16] load your warrior and all of the benchmarkwarriors, and then choose challenge by: [yourwarrior] [23:16] self-fights are not needed, preferably atleast 100 rounds per matchup [23:17] where do I find the benchmark warriors? [23:17] depends, you can use a ready-made benchmark, or make your own [23:17] www.koth.org/wilkies/ is one to start with [23:17] http://www.corewar.info/bench.htm [23:18] go with the wilkies, anything better will be more of a suicidal than depressing [23:19] and: http://www.corewar.info/optimax/dload/ [23:19] he doesnt really need optimax right now [23:21] 17.8 :D [23:24] maybe that explains why im so willing to tell beginner's how to improve their warriors [23:25] 35.4 with a core of 800 [23:25] the warriors are optimized for core of 8000 [23:25] for that there are other benchmarks [23:27] so I made a dwarf that changes the bomb distance when it hits the end of the core [23:28] "the end of the core" ? [23:28] the dwarf's original stepsize (4) is already problemous enough [23:28] so you might want to start with bigger step and then switch to smaller [23:28] distance of 8, changes to 7 after 1000 runs [23:32] http://pastebin.com/m389c136c [23:33] you have a bug.. [23:33] what does mov.ab $bomb, @bomb do? [23:33] what does add.i do? [23:34] (same as add.f) [23:35] adds bomb.a to bomb.b [23:36] wrong answer to both of our questions [23:36] I assume that's not the best way to do that? [23:36] er, that's the add line [23:37] mob.ab places a bomb at the address of bomb.b [23:37] no it doesnt [23:37] what does mov.ab do? [23:37] well, no i basicly moves a values into the a field [23:37] use "mov.i" [23:39] well, it does place bombs correctly [23:39] it places values, not bombs [23:39] aha! [23:39] use the core-view to see [23:40] I only debugged by myself, so it appeared the same [23:45] and i still dont understand why you use the add.i instead of add.ab what the original dwarf used [23:46] I wasn't looking at the original when I wrote it [23:46] ah [23:46] to facilitate learnin's [23:53] Part: Dwaddle left #corewars [23:55] my score went down to 15.9 D: [23:55] whoops, didn't save. [23:57] i get 53 from properly working version [23:59] 74! [23:59] woo!