[00:13] Part: Mannerisk left #corewars [00:20] MSG: Quit: humhum [00:26] Join: Mannerisk joined #corewars [00:41] MSG: Ping timeout: 240 seconds [01:57] MSG: Quit: flyduck [02:31] Join: flyduck joined #corewars [02:33] my newest warrior has a 40% win rate against rave :D [02:46] Join: Mannerisk joined #corewars [03:07] i remember rave :D [03:07] so long ago wow [03:07] i wonder how my warriors score against it [03:08] oho [03:08] 66% for re-bolo [03:08] lol [03:08] i don't know how to read these columns maybe that's backwards :( [03:10] i don't see how it can lose 80% to the thing i was messing with for the mini challenge [03:18] i just started with corewars 2 days ago and have been using it as a benchmark [03:18] i just did something cool as a test tho [03:18] my first replicator :D [03:18] :) [03:19] ;data [03:19] reloc EQU 100 [03:19] length EQU 6 [03:19] org start [03:19] start MOV.I 0, reloc [03:19] ADD.A #1, $start [03:19] ADD.AB #1, $start [03:19] jump DJN start, #length [03:19] ADD.AB #3, $(reloc-1) [03:19] JMP reloc-length+1 [03:19] :D [03:19] took some fiddling (sorry for the long paste) [03:19] lol my best scoring warrior vs rave is a broken test thing [03:20] D: [03:20] but bolo gets third place at 57% :D [03:20] my thing that scores 40% win against rave is just an .8c scanner and then a one-shot clear, very basic [03:21] and all code basically copy+pasted from HOWTOs, with me fiddling with it for understanding of course [03:22] also i started more than 2 days ago, that was a typo, but it was just last week :P [03:22] more like 5 days [03:22] oh nevermind [03:22] it's not broken i just couldn't see what i was doign [03:22] doing* [03:22] wow, lol [03:22] i basically unrolled a bomb loop and then booted a dwarf [03:22] 72% win [03:22] i lold [03:23] i still haven't written a scanner [03:27] lets see how that replicator thing i just did fares against rave... [03:27] even though it doesnt do any attacking, just replicates [03:27] not very good :-x [03:27] hehe yeah [03:27] it just runs around and waits to get hit [03:27] most of mine didn't fare well either [03:27] but bolo was the exception [03:27] which was interesting [03:28] unfortunately the self-healing version(s) didn't score as good as the original :( [03:28] i had a thing on my first warrior that tried to move itself when it saw it was getting attacked, but when i removed it my win-rate against rave jumped from 20 to 40 [03:29] it wasnt a very sophisticated defense though [03:29] hehe [03:29] you can't move fast enough once you are already getting bombed [03:29] but! [03:29] just a bunch of NOPs, a JMN 0, nops and then a bunch of mov's [03:29] you could try having a couple copies watch each other's backs [03:29] and nuke the hell out of one to avoid getting stunned [03:29] hehe [03:30] still in the end stuff like that doesn't really work well enough :( [03:30] i want to add an attack to that replicator i just did but dont really know where to go from there... [03:30] any ideas? [03:31] you could start by learning how silks work [03:31] what is that? [03:31] i don't know if there's any use to doing a loop style thing like tht [03:31] a silk uses stacked processes to copy itself [03:31] for example your first two commands might be [03:32] spl step, 0 [03:32] mov >-1, }-1 [03:32] you create one process for each line before running it [03:32] so if it's 8 lines long you can use three spl 1's beforehand [03:33] papers use that, two or three of them even [03:33] to make a ton of copies and lots of processes [03:34] on the other hand, if you want to work with the code you have [03:34] you can make the two adds into one command [03:34] even if it's just [03:34] nop }start, >start [03:35] :O [03:35] i didnt realize a NOP could have increments in it like that and it'd work [03:35] nice. [03:35] but mov/nop wouldn't be as good as just mov [03:35] for example you could have [03:35] ptr dat start, reloc [03:36] mov }ptr, >ptr [03:36] djn... [03:36] then you are incrementing and moving all in one command! :) [03:36] nice... [03:36] it starts to be less of a leap from that to the spl thing once you see the little optimizations inbetween [03:36] i knew there must be a way to shorten those adds but i couldnt figure it out [03:37] then instead of dat you use SPL to hold your two locations [03:37] thogh, that's making a copy [03:37] not "running away" [03:37] i wonder why random keypresses get lost on my laptop keyboard [03:38] hehe my favorite warrior so far is not very useful but it sure is complicated :D [03:38] it is a replicator too [03:38] it's 4 instructions in 2 different places in core, they synchronize with each other to play leapfrog [03:38] i can't really think of any ways to make it score good though lol [03:55] brb! [03:55] MSG: Remote host closed the connection [04:06] Join: flyduck joined #corewars [04:19] MSG: Ping timeout: 240 seconds [04:26] MSG: Quit: leaving [04:38] MSG: Ping timeout: 240 seconds [04:43] Join: Mizcu joined #corewars [05:23] nescience, whew [05:23] i just finished shortening it to 5 [05:23] there were a few tricks involved SLIGHTLY different than what you said, and it didnt help that im tired [05:24] ptr DAT ptr, reloc [05:24] start MOV }ptr, >ptr [05:24] jump DJN start, #length [05:24] ADD.AB #(jump-start+1),$(reloc-1) [05:24] JMP reloc-length+2 [05:24] : ) [05:25] actually change that to JMP reloc+start, it [05:25] 's smarter.. [05:28] haha wow... [05:29] this is a lot better than the last one because by adding the dat line it makes for an attack of sorts... [05:31] lol [05:31] good job [05:31] Nick Change: myndzi\ changed nick to myndzi [05:31] :D [05:32] now add an attack too! [05:32] i just ran it against rave and 8 matches took a really long time on -v 4 [05:32] i don't know what -v is [05:32] :\ [05:32] its the fastest you can run the simulation in pmars, i think [05:33] or wait, slowest, woops [05:33] ha [05:33] -v 903 or 904 is slowest [05:33] yeah [05:33] o rly? [05:33] you must have a really old computer? [05:34] even my laptop doesn't take "a really long time" to run 8 matches [05:34] slowest in terms of pmars, not in terms of practise [05:34] :\ [05:34] use corewin if you need slowness [05:34] lol. [05:35] im using pmars in linux, it does use 99% of my 1st cpu sometimes :o [05:35] but its not that old of a laptop [05:35] pmars uses all the power it can get [05:37] for a single processor, that is [05:37] there is no multicore-mars [05:37] wow, that seems to beat rave some of the time : O [05:37] i just went 8 - 6 - 2 with it ! [05:37] 8 wins? [05:38] yeah [05:38] o_O [05:38] hehe [05:38] with only one process and no attack to speak of i am a bit surprised [05:38] maybe it's just lucky random numbers [05:38] :P [05:38] i think so :P [05:38] i just did it again and it was 6 and 9 [05:38] ill do 1024 matches and paste the result.. [05:39] that's why you usually run like 200 matches [05:39] its interesting to me how much offsets affect things too [05:39] if the step is at 100 its barely effective but if you up it to 101 it becomes much more [05:39] yeah, i like pretty numbers [05:40] of course since you are testing it against only one warrior, there are gonna be some numbers that are way better [05:46] 422 - 433 - 159 [05:47] about 45% / 45% / 10% [05:47] :o [05:54] weird [05:56] i wonder if it'd have any chance on a nano hill, it is 5 lines [06:17] it'd get destroyed [06:18] i recommend submitting it for educational purposes ;) [06:18] hehe [06:33] Part: NorthStar left #corewars [07:12] ok i gave up trying to make it 4 lines by cramming a number into the b-field of the JMP [07:12] too tired [07:12] but i suspect it could work... [07:15] Join: Metcalf joined #corewars [07:32] you could technically make it like two lines [07:33] if you want to let a couple processes fall off the end :P [07:33] and maybe you could do it without that too [08:14] MSG: Read error: Operation timed out [08:15] MSG: Read error: Operation timed out [08:19] Join: flyduck joined #corewars [08:30] Join: bvowk joined #corewars [08:47] MSG: Read error: Operation timed out [09:16] Join: sh0ne joined #corewars [09:16] hello [11:04] MSG: Quit: Leaving [11:09] Join: sh0ne joined #corewars [11:27] MSG: Quit: Leaving [11:47] Join: sh0ne joined #corewars [13:05] MSG: Ping timeout: 240 seconds [14:05] Join: fiveop joined #corewars [16:14] hrm. [16:22] Hmmm...? [16:23] Join: flyduck joined #corewars [16:24] hey met. [16:24] Hacking anything good? [16:44] No hacking today [16:44] Just some tedious html stuff [17:46] Join: Fizmo joined #corewars [17:47] hi hi [17:52] MSG: Read error: Operation timed out [18:07] Join: flyduck joined #corewars [18:08] Hi Fizmo :-) [18:31] MSG: Read error: Operation timed out [18:43] Join: flyduck joined #corewars [18:52] Join: Fizmo_ joined #corewars [18:55] MSG: Ping timeout: 240 seconds [18:55] Nick Change: Fizmo_ changed nick to Fizmo [19:03] MSG: Quit: ChatZilla 0.9.84 [Firefox 3.0.4/2008102920] [19:06] MSG: Read error: Operation timed out [19:11] Join: flyduck joined #corewars [19:16] Join: NorthStar joined #corewars [19:39] MSG: Read error: Operation timed out [19:47] Join: elkauka joined #corewars [20:09] Join: flyduck joined #corewars [20:29] Heapsort in Redcode still takes up about 60 instructions [20:35] MSG: Quit: blaaaarghhh [20:48] * Metcalf starts writing a quine in Redcode [21:03] Metcalf, i did that last night :D [21:12] Pfff, to late again! [21:12] I've only just found a pen [21:15] MSG: Quit: humhum [21:57] MSG: Ping timeout: 240 seconds [22:14] Argh, how do I set maxlength on the exMARS command line? [22:49] Quine is about 160 instructions, trying something different [23:22] MSG: Ping timeout: 240 seconds [23:30] Alternative quine is 215 instructions :-(