[01:22] Join: Caelian_ joined #corewars [01:26] MSG: Ping timeout: 240 seconds [04:43] Join: Dirty|120 joined #corewars [04:43] Nick Change: Dirty|120 changed nick to Luka [04:44] Hi everyone :) [04:45] * Luka is new here [04:51] noone seems around, but I feel like talking alone :P [04:52] So, I'm new here, but not 100% new to redcode and warrior programming (very) basics [04:52] never got skilled enough to enter a hill tho (quite far from that) [04:53] I'm wondering if someone got some easier benchmarks than the three classical wilXXXs [04:54] if not, as a learning course I'm building some, from trivial to ... how good I'll get [04:55] probably still weak then hehe [04:56] but I bet a lot of beginners are giving up early, and I thought it mays help to have something easy to beat at the start [05:00] hi Luka [05:01] well, i had been thinking of putting one beginner-mark up recently, but never got up to it [05:02] http://koti.mbnet.fi/mizcu/bench.txt <- there is what i had put up [05:02] Hi Mizcu :) [05:02] having a look right now [05:03] almost all the warriors in that list can be found at Koeningstuhl [05:03] Koeningstuhl? :) I'm affraid I'm not used to all the good places yet ;) [05:04] http://www.ociw.edu/~birk/COREWAR/koenigstuhl.html [05:05] and I think what I'm building atm are beginner-mark from very very very very easy to very easy according to your ranking hehe [05:06] dont be decieved - for a beginner even that first one can give headaches [05:07] and that "medium" will have some that give easy scores, and few that at first seem impossible to beat [05:07] Wilfiz is equivalent to average-good (imho) [05:07] the scores one gets out of wilfiz are about 10-20% less that at the KOTH-hills [05:08] I'm quite far to score something descent with wilfiz ;) [05:08] I see [05:08] and then there is the Fshnop0.3 , or just 'nop0.3 which is quite close to koth-hills [05:09] well, building a warrior is a learning process [05:10] it takes time to learn to see where there could be improvement and where not [05:12] Want to have a look to my last warrior? so you'll have an idea of how I'm doing atm [05:14] I sure can [05:15] should I send you a mail or just paste the instructions in private? (it's a rather short one) [05:16] if its under 5 lines, just paste it here, under 10 lines, straight to me, if more then try pastebin (pastebin.com) [05:18] 9 lines ;) pasted in private [05:19] first of all, you only need one spl [05:19] a little trick which not obvious at first: spl $0 and spl #0 act the same [05:20] but it has to do with the way the codesimulator works [05:21] spl $4, of course splits a process [05:21] when the codesimulator starts runnin the spl, it checks the target where the splitting goes [05:21] $4 gives the current+4 as the position [05:22] however, #4 is wrong datatype for position. In '88 it was not allowed to be used, but in '94 standard it is now allowed [05:22] when the spl with "wrong" datatype is run, the target where the splitting is done defaults to 0 [05:22] I see [05:22] so spl #666 acts like spl $0, but you can store an extra number in it [05:23] I got the point but [05:23] "wouldnt that cripple the coreclear?" [05:23] I need to store 2 numbers in the first spl [05:23] hehehe [05:24] spl #0, #4001 is perfectly viable [05:24] in fact, spl #0, $4001 is fine too [05:24] combining the two spl's would lead that the mov -instruction in the coreclear is no longer modified [05:24] I like spl $0 for another reason, since it changes and become spl $5, spl $10 and so on, it stunes my opponentbetter [05:25] stuns* [05:25] but you can put the oh right [05:26] such spl $5 (4 dats) spl $10 (4 dats) spl $15 (4 dats) doesnt seem that effective to me [05:26] you might want to compare the choices on a benchmark [05:29] I'm trying the changes you suggest right now [05:33] 3 more points against wilfiz ;) [05:40] but now this basic warriors seems just weak and hard to tweak without adding something significant ;) [05:43] your stepsize for bombing is currently 5 [05:43] change that to something bigger, not just 15 or 25, but say, 2055 [05:44] I'll have a trouble here [05:44] I use a add.f #5, bomb [05:44] theoretically, the best bombing pattern is where each bomb "divides" the core into smaller and smaller piece. However, there is no such step. [05:45] yeah [05:45] but how to do about the add.f trick? [05:46] you could move the bomb away from your warrior [05:46] say bomb equ start+step [05:46] relative address bomb is 4, I use a small increment to make it 5 [05:46] and then use add.f #bomb, bomb (that will work, believe or not) [05:47] (and mov.i splinstruction, *bomb etc) [05:48] I have troubles to see how that works [05:49] the idea is that the mov-instruction adds the numbers to an instruction far away from you [05:50] and you still use the far-away instruction as pointer for the bombs [05:50] oh! [05:50] and add.f #step, step will in turn make a dat step, step at (step) away [05:52] and will it works with the jmn too? [05:52] that is something i am slightly unsure of :) [05:52] hehe [05:52] you might have to change the trigger to coreclear because of it [05:53] but djn.b dwarf, #1300 is idiotically sure to work [05:54] er not as fun, but true as I don't need the >bomb anymore [05:57] ok trying to have all that working :P Not sure about the new add.f (and everythin that goes along) yet ;) [05:58] just give it some time [05:59] there is a small preference not to give out straight code to beginners [05:59] I can understand why [06:04] brb, gotta take dog for a walk [06:05] see ya later :) hope I'll have it working while you are out hehe [06:05] thanks already for the help [06:16] back [06:17] not doing too bad, except that I can't get my two mov.i aiming at separate addresses ;) [06:20] add.f #-step, step [06:23] doesn't seem to work, but trying to figure by myself ;) [06:33] I'm bombing myself everytime :P [06:54] Join: CrisisSDK joined #corewars [06:55] .. [06:55] Hi Crisis [06:55] Hello [06:57] Hi Crisis ;D (idiotic beginner here) [06:59] How may i be of assistance? [07:01] you did enough for me already ;) trying to understand why I needed those tricks while your solution seemed simplier [07:02] well I'm a complete noob at corewars [07:05] So am I, or not far ahead :P [07:07] Crisis: well, then you could start by taking a deep look at www.corewar.info [07:08] You'll probably be further then me... It's different from Binary Armageddon.. [07:08] you from Brazil? [07:09] no wait, it was Memory Armagedon, disregard that question [07:12] ok. (Binary Armageddon was made by NecroBones... ) [07:15] well, for any questions you have, just ask here, someone will answer, and even if not available, probably will at the logs [07:16] ( www.koth.org/irc-logs/ ) [07:19] ok thanks [07:30] If the best warrior is the imp, the second best one must be JMP 0 [07:31] not really [07:36] yeah, imp isn't the best warrior [07:37] and JMP 0 just keeps hoping that the other warrior will kill itself [07:42] score 78 on the no P-space hill ... still a long way to go ;) Thanks Mizcu for the various helps, off to bed myself :) [07:42] scored* [07:44] I think I might make a replicator [07:50] ... [07:51] Can I have a MOV statement move a command referenced by one DAT to a position referenced by another DAT ? [07:52] mov.i *reference, *reference [07:53] yes [07:53] ok [07:53] so a * [07:53] @ will do too [07:53] depends on where the reference is [07:53] is the reference on a- or b-field of dat [07:53] * uses a-field, @ uses b-field [07:54] ok [07:54] aI'll check [07:56] mov.i @newloc @memloc isn't working, and just to make sure I made the DAT statements to be like: DAT 10,10 [07:59] hmm [07:59] want to give the code so i can get a look? [08:05] ok, I just copied straight from BA by the way [08:06] hmm... [08:06] I think I'll put it all on 1 line so I can post it... [08:07] jmp 4 :memloc data -1, -1 :newloc data 25, 25:numsects data 12, 12:main do @numsects mov.i @newloc @memloc add @memloc #1 add @newloc #1 loop -3 mov @memloc -1 add @newloc 10 jmp main [08:10] ugh, thats an annoying thing to read [08:11] I agree [08:11] I tried getting rid of all the labels, but it still didn't want to work [08:12] let me guess, do - loop runs the stuff between for (numsects) times? [08:12] not familiar with BA so cant be sure [08:13] yes [08:14] by pulling the thing apart in various different ways, I think I can safely say that the problem is something to do with the MOV [08:14] mainly because thats where it dies [08:15] i dont know how the references are done in BA, but you seem to overuse them [08:16] add @memloc #1 will add to the place pointed to by memloc, not in the memloc itself [08:17] in BA that it is reccomended you use the :label thingys like that, calls it psudo variables or something [08:20] in BA instructions work as mov target, source? [08:25] either [08:25] you can set it as that or mov source, target [08:26] because this is what i think you are after in the warrior [08:26] jmp 4 [08:26] data -1 [08:26] data 25 [08:26] data 12 [08:26] do @-1 [08:26] mov.i @-4, @-3 [08:26] add -5 #1 [08:26] add -5 #1 [08:26] loop -3 [08:26] mov -8, #-1 [08:26] add -8, #10 [08:26] jmp -6 [08:28] ok, I'll try that [08:28] yep [08:28] only it still dies at the mov [08:32] Join: Roy joined #corewars [08:32] Nice Corewarrior guys! [08:32] h R [08:33] * Roy is very curious about the bvokium and the vamp part, don't have time to read it now though [08:33] Ohw, and the reversed engineered Blacken is funny, I forgot all about that... its was a painstaking process [08:35] But well, I'm going away for a long weekend, bye guys! [08:35] MSG: Client Quit [09:35] Part: CrisisSDK left #corewars [10:03] :) [10:07] ^ [10:07] I see that you've been quite busy tutoring newbies, Mizcu :) [10:08] duty called [10:08] hehe [10:39] It seems as if I've finally managed to post #92 without having it mangled. [10:39] * Fluffy kicks Google again [11:15] Hello :) [11:28] Join: elkauka joined #corewars [11:37] hi@all [11:40] Hi Elkauka :) [11:41] someone can get pastebin.com to work? [11:46] MSG: Read error: Operation timed out [11:47] Join: Fluffy joined #corewars [11:47] http://fr.pastebin.ca/492732 [11:55] Nick Change: elkauka changed nick to elk|afk [12:25] MSG: Quit: and over there we have the labyrinth guards. one always lies, one tells the truth, and one stabs people who ask t [12:59] Join: fiveop joined #corewars [13:10] Looking for the more possible basic example of scanner, if anyone knows where to find it :) [13:27] Join: Core_old joined #corewars [15:00] just wrote a cmp-scanner from scratch, would appreciate some basic comments ;) (no constant tweaking yet) [15:00] http://fr.pastebin.ca/492962 [15:00] er [15:00] wrong link [15:01] http://fr.pastebin.ca/492968 [15:01] better :) [15:02] updated, was still wrong, oh well -_- [15:05] Join: elk|afk joined #corewars [15:05] Nick Change: elk|afk changed nick to elkauka [15:08] tobomb mov start, cmp.ab scan, scan [15:08] jmp tobomb [15:08] add.ab #gap, scan [15:08] jmp update, ooops [15:08] sorry [15:11] ok this time it should definitly be ok: http://fr.pastebin.ca/492987 [15:11] if anyone feels like droping a comment [15:11] to help a newcommer ;) [15:44] the second cmp doesnt do what you are wanting it to [15:44] oh? seems to work fine when I run it tho, what's wrong with it? [15:45] hmm, well, i am just thinking too sophisticated [15:46] it could do something more? :) [15:46] it gave me an idea, but it would be unefficient [15:48] ok :) I did a change myself, I'm using spl 0 bombs rather than dat ones, since the coreclear is working fine [15:50] well, other than that i see nothing that immediately makes me suspicious [15:51] nice, thanks :) [15:51] with switching tobomb and cc -parts you could get rid of one instruction, but itll be harder to switch to cc [15:52] and the tobomb could use one change [15:53] using sne instead of cmp? [15:54] no, using mov.i bomb, doesnt save space, but it throws one bomb per two cycles, instead of one per three [15:55] I meant to switch cc and tobomb parts :) [15:56] ah, quite so [15:56] though you might want to teach yourself to use seq instead of cmp [15:56] does the same? [15:57] yes [16:03] ill go now into semi-afk situation [16:03] might be back before its dark, might not [16:03] ok, Mizcu, thanks again :) [16:56] Join: John joined #corewars [16:57] Hi :-) [16:58] hi john [16:58] Hi El Kauka [16:58] Nice article :-) [16:58] thanks ;) [17:00] hopefully tons of vampirs are going to storm the hills now ;) [17:00] * John goes to download the new version of Corewin :-) [17:01] he seems to have changed it. the link point to the old version [17:02] i've downloaded ot early this day. backup here: http://rapidshare.com/files/31842722/CoreWin.exe.html [17:02] Hi John :) [17:02] Hi Luka [17:03] Elkauka: what's old? It says 2.3 on the page [17:03] but the exe is the old version [17:03] Are you new here Luka? Or have you been here before with a different nick? [17:04] new :) [17:04] I'm making sure to bug everyone a little since a few hours to learn some tips hehe :P [17:05] Oh :-) Welcome then... [17:06] Elkauka: I downloaded it and it says v2.3 in the about box [17:07] Have you sent anything to the beginner hill yet Luka? [17:07] weired. a few minutes ago i didnt work [17:08] well nevermind then [17:09] Join: Fizmo joined #corewars [17:09] hi hi [17:10] hi fizzer ;) [17:10] hi John, hi el kauka [17:10] Hi Christian [17:11] very interesting cw issue :) [17:11] I'm trying John :) [17:11] hi Fizmo :) [17:14] hi Luka [17:16] Join: Fizmo_ joined #corewars [17:18] hmmm affraid I'm submited a warrior twice, including once as anonymous :-/ is there a way for me to kill this one? [17:18] I* [17:18] MSG: Ping timeout: 240 seconds [17:18] on which hill? [17:18] Nick Change: Fizmo_ changed nick to Fizmo [17:18] SAL, 94b [17:19] send a dat 0, 0 [17:20] and add this line [17:20] ;kill _____ [17:20] while ____ is the name of your warrior [17:21] Or just send: @kill 94b warriorname [17:21] Just one line, and it doesn't leave a dat 0,0 warrior on the hill [17:22] first option doesn't seem to work gonna try the latter ;) [17:23] It killed the correct one too :P [17:23] but it's alright [17:23] gonna tweak it a bit before submitting again [17:26] ;-) [17:31] Join: Fizmo_ joined #corewars [17:33] MSG: Ping timeout: 240 seconds [17:33] Nick Change: Fizmo_ changed nick to Fizmo [17:36] let's see how good it does now ;) [17:36] er [17:36] I meant how bad :P [17:37] hehehe [17:38] 19 31.0 55.5 13.4 FireWorks Luka 106.5 [17:38] yay! :D [17:40] neat! [17:41] Nice :-) [17:41] thanks for the help El Kauka :) [17:42] the D-Clear helps much :P in fact it does most of the job, not that deserved for me to get on the hill lol [17:42] http://fr.pastebin.ca/493234 [17:52] :) [17:56] * Fluffy congrats Luka for SmallNNasty and FireWorks :) [17:56] Hi Fluffy [17:56] Join: Fizmo_ joined #corewars [17:56] Hi Fizmo [17:57] @Fluffy Thanks a lot :) [17:58] MSG: Ping timeout: 240 seconds [17:58] Nick Change: Fizmo_ changed nick to Fizmo [18:01] hi Fluffy :-P [18:01] how r u? [18:02] I hope to survive the day :) [18:02] hehehe, why? [18:02] All idiots seem to have chosen this day to show how stupid they can be [18:03] lol [18:03] btw [18:03] did anyone tried to send something on koth server [18:03] I sent something but it seems not to arrive [18:05] I've just sent a DAT and now it is in the queue [18:06] I sent some today [18:06] they did bad tho ;) [18:06] bad? [18:07] It seems as if you've sent 4 warriors and now you are already on the hill, Luka :) [18:08] hmmm [18:08] maybe my will appear sometimes [18:09] It seems as if Google finally has managed to show up #92 without mangling it. [18:09] * Fluffy kicks Google [18:09] Fizmo: the hill now shows my warrior [18:10] hmm [18:10] Fluffy: On SAL's beginner hill ;) [18:10] Luka: yes. What's wrong with it? :) [18:13] Fluffy: nothing ;) my warriors did bad on KOTH server tho :) [18:14] Luka, you can't expect to enter the '94nop-hill right after starting with cw [18:14] It usually takes some time :) [18:15] hehehe [18:16] well I guess I should better get back to work on those other warrior types I didn't understand yet ;) [18:16] If you want some inspiration try to read some old issue of the Core Warrior [18:17] I'd like to read the wiki too, but seems down :-/ [18:17] Luka: there's an index here to Corewarrior http://corewar.co.uk/biblio.htm [18:18] Unfortunately the Wiki has disappeared. [18:18] old issues of the core warrior are a bit tough for me still, but will try [18:18] Hi John :) [18:18] with all its datas? :-/ [18:18] Yes, the data has gone forever :-( [18:19] I tried core war first like 2 years ago, didn't invested much but remember reading some nice articles about imp-spiral launchers and quick scan on the wiki ... or was it in core warrior? [18:20] Luka: There are some posts on rgc about how quickscanners work [18:20] Simply go to http://groups.google.com/group/rec.games.corewar and searcher for Quickscanners [18:21] ok, time to leave for me [18:21] * Fluffy waves [18:21] * Fizmo waves [18:21] * John waves [18:21] c ya [18:22] see ya Fizmo :) [18:22] Fluffy: ok will have a look [18:22] elkauka: Have you already noticed that vamps are back is on Koenigstuhl :) [18:22] MSG: Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.3/2007030919] [18:23] * John goes to look [18:23] 51st on 94nop :) [18:24] yeah, have looked it up already [18:27] :) [18:27] I really hope to see more vampires soon [18:28] might have a look hehe [18:28] have to disconnect, brb [18:28] * Fluffy waves [18:28] MSG: [18:29] Time for me to go :-( [18:29] * John waves [18:29] * Fluffy waves [18:30] MSG: Quit: mov.i #1,1 [18:32] Join: Luka joined #corewars [18:32] welcome back [18:32] thanks :) Can you give me the link to corewarrior issues again? :) [18:33] http://www.corewar.co.uk/cw/index.htm [18:33] thanks :) [18:33] http://www.corewar.co.uk/94war/index.htm is even a little bit older [18:33] And http://www.corewar.co.uk/pushoff/index.htm is acient [18:33] *ancient [18:34] hmmm I think the link you gave me earlier was different, issues were sorted by topic [18:34] http://corewar.co.uk/biblio.htm [18:35] yes thanks :) that's just great [18:36] Have fun [18:36] I'm having much :) [18:41] Luka: Don't hesistate to ask, if you have questions :) [18:44] damn it. they took down my ftp backup server a few days ago. my first brainf*ck compiler - gone forever :'-( [18:45] Then use another brainfuck compiler :) [18:45] :P [18:52] and shouldn't you write a successor to vamps are back, elkauka? :) [18:56] fizmo wanted me to rewrite it for '88 [18:57] any success? [18:57] havent started yet [19:21] c ya folks [19:21] MSG: Quit: and over there we have the labyrinth guards. one always lies, one tells the truth, and one stabs people who ask t [19:21] * Fluffy waves too late [19:24] to you, what seems the easier warrior concept to understand: QScan or Silk? [19:26] hmm [19:27] early quickscanners are quire easy and then you only have to make little steps until you will reach the current q^4.5-scanners [19:27] *quite [19:27] gonna try that then :) [19:27] and I would say that they are a little bit more simple than papers [19:28] In any case it is hard to really tell you that now that I already know how both work [19:28] http://www.corewar.info/lexicon/paper.htm is the text that I've used to learn all about papers [19:29] thanks for the link :) [19:40] very nice reading :) [19:43] yes, how we can expect from you "very nice" warriors :) [19:44] to come up with something original will take a while :P [19:47] First it helps to try out the "old" strategies [19:49] I already built some simple warriors using those "old" ones, I bet I missed some, but have like 10 different small warriors [19:49] they are quite usefull as a beginner-mark ;) [19:49] Did you benchmark you warriors? [19:49] yep [19:49] agianst Wilkies? [19:49] *against [19:50] I'm doing poor against those wilXXXs benchmarks [19:50] but some of them always do well against some of them, since most of them have a nemesis [19:50] that's normal [19:51] you have to understand that there are certain things that you can usually take for granted [19:51] for example paper usually win against stones [19:51] stones win against scanners [19:51] and scanners win against papers [19:52] at least that's true for "old" warriors [19:52] So a good benchmark will always contain warriors, which win without problems against your warrior [19:56] yep :) that was the underlying idea in the benchmark I'm building too [19:57] but the warrior are way simplier [19:57] helps me to see how good my new warrior is doing against D-Clear, IMP-spiral and so on [19:58] I see [20:00] Join: DirtyBria joined #corewars [20:00] MSG: Read error: Connection reset by peer [20:00] Nick Change: DirtyBria changed nick to Luka [20:07] Nick Change: Luka changed nick to Dirty|afk [20:48] MSG: Quit: Trillian (http://www.ceruleanstudios.com [22:33] MSG: Ping timeout: 240 seconds