[00:53] Join: Mannerisk joined #corewars [07:35] MSG: Remote host closed the connection [07:48] Join: myndzi joined #corewars [09:03] Join: zebs joined #corewars [09:06] Hey guys. Im having trouble subtracting a number from the b-field with another number which is contained in the b-field of another DAT. I could not get this to work as i wanted, so i moved on to trying to save one of the numbers to the private data storage. And then adding one of them to the a field of the other, thus subtracting [09:08] This is not my actual warrior but just trying to solve how to manage wiith the numbers: http://pastebin.com/m34524def [09:08] I want so subtract 120 from 140 and get the answer of 20 [10:46] dat 0, 5 [10:46] dat 0, 4 [10:46] sub.b -1, -2 [16:06] Mizcu: http://pastebin.com/m45c918f8 [16:07] I want the calculation described in the paste to be done. Cant get it to work :/ [16:09] sub.b 2, -1 [16:10] if you want to use indirect addressing, sub.b @-2, -1 [16:10] ah, I thought that you couldnt use the @ in the a-field ? [16:10] And that u should use * instead? [16:11] you choose * and @ depending on which field you want to use in the target instruction [16:11] it doesnt matter on the running instruction [16:11] ah :) [16:11] Cant believe I've been stuck on such a shitty detail :D [16:11] thx mate [16:11] works like a charm now :) [16:14] correction: you choose depending on the field you want to use in the pointer-instruction [16:15] target-instruction would depend on the modifier (.ab) [16:18] yepp, I actually read it like that haha ;) [16:20] Hi :-) [16:20] h [16:48] Mizcu: here is my warrior. :) It does very well against dwarfs and the like, but does not perform good against imps. What is a good strategy to dispose of imps? Scanning for mov 0,1 and moving rapidly forward until i wont find any more and then finish him off with a dat? [16:49] Also, im using corewin and tournament mode with 3 warriors. My warrior wins the majority of matches by far but scores the least in the summary. Very weird :/ [16:54] ah no matter.. :) [17:55] Join: fiveop joined #corewars [18:32] back [18:34] best method to fight against imps is to keep constantly modifying a position behind you [18:35] eg. you can turn a boring looping instruction (jmp -4 etc) into an jmp -4, <-10 which will help a little [18:47] This is actually supposed to be a school assignment. [18:47] To defeat an imp and a dwarf [18:49] i dont recall if it was you, but i remember a previous student appearing here [18:49] i think it was a previous year's student [18:49] Yeah, one year ago, he is a friend of mine and he adviced me to go here :D [18:52] I do scan for modified positions in the core and then drop a bomb on the adjacent positions to the link. This will kill a dwarf but not an imp so after that i just keep dropping bombs startig from behind of my program and decounting. Killing an imp, if it hasnt already overwritten me while trying to find modified spots [18:55] what's the deadline for the assingment? [18:55] I thought of a self-replicating program, would that work? Make a backup of the program some positions ahead and then doing a comparison to check that they are untouched [18:55] Wednesday [18:55] it easy as long as you dont have to fight them both at once [18:56] yepp [18:56] as i said, you should fight by modifying a location behind the warrior [18:56] your scanning-loop probably modifies a jmz or cmp/seq/sne to find opponents [18:56] use an indirect scan, and set the pointers behind the warrior [18:57] hmm :) [18:58] thats should be good enough without giving any actual code [18:58] yeah, but why pointers behind the warrior? to compress it? [18:59] because the pointers keep changing every once in a while [18:59] if an imp overwrites the pointer, it might get changed, and thus damaged [18:59] aha okay [19:00] and if you are fighting against only an imp, it doesnt matter if it screws up your pointer, if you kill the imp in exchange [19:00] you mean if i run a cmd using the pointer thus changing one or more of the fields for the imp ?? [19:01] if you use a command that changes the pointer [19:01] gotcha [19:02] "but theres only one change of pointer in the whole loop!" , but all instructions have the capability to manipulate it somehow [19:03] We are going to use a tournament setting to compete (i guess we will be about 15 ppl). The prerequisites to compete is that you win the majority against imps and dwarves [19:04] Hence i think that most ppl will try to build a warrior thats good against them. Is there any kind of strategy you can think of, just like that, that maybe would be good?` [19:05] i guess that as the only one of your group here in irc, you have advantage [19:05] i suspect that most of the warriors will be jokes or existing warriors with very little changes.. [19:05] a dat-bombing scanner should be fine [19:06] although if you suspect that someone will build a replicator, then itll get harder [19:07] replicator is someone that heals? or moves around? [19:07] something that makes new copies of itself [19:07] a okkay, spl [19:13] any increments for scanning that is particularly awesome? Using primenumbers or the like to prevent bombing the same spaces over and over? [19:13] Checking not bombing ofc [19:15] the scanning-pattern is actually quite dependant on the scanner [19:15] you want it the pattern to be as tight as possible without hitting/scanning yourself [19:16] in scanning, it is possible to make yourself "transparent" to your own scan [19:16] how big your warrior currently is? [19:17] 25 [19:17] thats a liiiitle too big [19:17] ;) [19:17] Yeah, but why not using up all the hundred? [19:18] the bigger you are, the easier you are to find and attack [19:18] (decoys not withstanding) [19:19] But using a dat #0, 0# to lengthen your program and then adding some misleading stuff at both ens? [19:19] *ends [19:19] dat $0, $0's [19:19] yeah. meants that :) [19:20] dat #0, #0, and $0 $0 will not be different to cmp in '88 standard, but they will be to seq.i/sne.i in '94 [19:20] we are using 94 [19:20] $0,$0 is default then?? [19:21] yes [19:37] Hmmm... I think for my next project I might emulate the PDP-8 in redcode ;-) [20:08] MSG: Ping timeout: 240 seconds [20:13] Join: nescience joined #corewars [21:37] Can anyone recommend a program for writing MPEGs to DVD? [21:38] Nero? [21:38] Mizcu: thx for all the help man. Time to go to bed [21:39] * impomatic waves [21:39] Nero's not free though :-) [21:40] you can get one for 40e, and you get a dvd-writer with it [21:41] well, i dont know if the express-version supports authoring [21:41] only vcd/svcd/minidvd =/ [21:43] MSG: [21:48] if you are using Windows Media or Ultimate there is a Windows DVD Author program [21:48] WinXP unfortunately, thanks :-) [21:48] I haven't seen many other programs that write to something a DVD player can read [21:48] what do you mean "unfortunately" [21:49] I'd try http://www.dvdflick.net/ [21:50] Well, I guess it's more practical than Vista, but it doesn't really help when I want to write a DVD. I've got Pinnacle Studio, but the resulting DVD was jerky [21:50] Thanks, will take a look now [21:50] Pinnacle is a hot mess of a program [21:53] :-) [21:54] DVD Flick looks like it'll do the job. I'll let you know how I get on :-) [22:27] MSG: Quit: humhum [22:28] i was just trying to do something like that the other day [22:28] i ended up using dvd styler [22:29] i needed to convert some h264 mp4s to a dvd video [22:29] worked well, but it didn't keep them to scale [22:29] i.e. it stretched to fit and i couldn't find any option to prevent that [22:29] if your video is already in the correct resolution though it should be fine? [22:31] Not sure, my girlfriend made the video. :-) [22:31] all i wanted was a 1-click solution to get all these demo captures onto a dvd player [22:31] it's surprising how hard that is to do, even now, nearing the end of dvd's lifetime [22:39] * impomatic is currently trying to figure out how to make a simple programming game using the URISC one instruction computer :-) [22:53] this is so typical.. [22:55] im in middle of my annual studying-angst, and now a local newspaper gives out an article that says that its a bad time to go from graduation to work, go study some more [23:00] :-/ [23:01] of course, it doesnt help that after maybe possibly getting a BEng, MEng requires 3 years of work-experience inbetween studies [23:01] *headslap*