[00:16] MSG: Quit: humhum [00:35] Join: Ussr1943 joined #corewars [00:35] MSG: Client Quit [04:39] MSG: Read error: Operation timed out [04:39] Join: crem joined #corewars [04:39] Join: bvowk joined #corewars [04:40] MSG: Ping timeout: 245 seconds [04:40] MSG: Ping timeout: 245 seconds [04:40] Join: zyph joined #corewars [06:18] MSG: Read error: Operation timed out [06:18] MSG: Ping timeout: 245 seconds [06:20] MSG: Read error: Operation timed out [06:20] Join: pak21 joined #corewars [06:43] Join: zyph joined #corewars [06:48] MSG: Read error: Connection reset by peer [06:49] Join: crem joined #corewars [07:49] Join: Nenad joined #corewars [08:03] Join: vanrijn01 joined #corewars [08:04] Nick Change: vanrijn01 changed nick to yoR [08:11] * Nenad greets Yor [08:11] :P [09:18] Hiya [09:41] MSG: Read error: Connection reset by peer [12:17] Join: fiveop joined #corewars [12:47] MSG: Ping timeout: 245 seconds [14:14] MSG: [14:58] Join: Kalle joined #corewars [15:01] hi [15:03] hello [15:03] how may i be of assistance? [15:06] Looking for a way to optimize my scanner so that I find the warrior before his clones kill me. It's a program to counter a specific warrior [15:08] what kind of scan-loop do you use, add/sne/jmp , add/jmz? [15:08] think you can take a quick look Mizcu? We are writing a counter for a specific warrior for class in school(Yeah I know redcode in school is kinda funny) [15:08] good question [15:08] let me pastebin it [15:08] well that explains all the newbies around here recently [15:08] Yeah we are from the same class =) [15:09] now i just have to figure out how i can get my school to throw a course on redcode [15:09] heck, i could be the teacher in it.. [15:09] It's a rather nice way to learn the thinking of assembly [15:09] http://p.cout.se/51 [15:10] first one is the warrior we try to counter, 2nd is mine. Problem is that his clones find me and then im dead. Wins about 53% of the time tho [15:12] a) your scan-loop sucks b) you drop only one stun-bomb before starting the coreclear [15:12] c) I'm a total newb at this? :P [15:12] I know it sucks [15:12] http://www.corewar.co.uk/scanner.htm [15:12] that should help [15:13] Well my warrior will drop a spl 0 and jmp -1 making his warrior spam new processes and become so slowed that he is useless [15:14] yes, it drops 'a' s/j [15:15] real scanners might drops hundreds during a run [15:15] That means I'll have to find his clones(which are ~200 steps ahead) and s/j them also? [15:16] it depends on how many clones he is running [15:17] well that depends on how long into the round we are. He keeps on splitting 200 steps ahead of the last one all the time [15:18] since you use a linear scan, you scan one instruction per two cycles, meaning that a worst-case scenario is about... (~37 copies = 14000 cycles..) [15:19] change your scanloop into something different and you can cut that worse-case scenario by about 5-10 times [15:19] which, even though not correct, can be said to correlate about 5-10 times more effect from a bomb [15:32] I wonder if I can implement a .8c scanner or if that wont be effective since I got less than 16 instructions alrdy [15:35] /whois Kalle [15:35] hello zyphy [15:35] hello [15:45] Kalle: you are not fighting against a 'real' opponent so go head with the .8c [15:46] one more thing Mizcu, a warrior that beats both an imp and a dwarf? [15:47] too many to mention [15:47] hmm, what would be simple enough example [15:49] and imp ring (or a spiral) would be one [15:49] though they look very voodoo before you understand how they work [15:51] with binary launch and stuff right? [15:57] yeah [15:58] Will look into it after I manage the .8c scanner=) thanks [15:58] ill be here for all day, so just ask if you have something [15:59] Shouldn't say that. Prolly will be back to bug you :p [16:01] its not like i go do a brain surgeon meanwhile ( http://www.mobygames.com/game/life-death-2-the-brain ;) [16:24] Mizcu: mind explaining how the .8c works? I get that it compares 2 places x2 but how can I use my target variable and still get the 4/5 speeD? [16:25] you cant use the target variable [16:25] but its not really required [16:26] because I wanna find his main program and not the clones. Wouldnt this scanner find anything? [16:26] well, that would need some changes in the scan-loop [16:27] Hmm I'll just see if I can find something alittle faster. Alrdy done just wanted to optimize it alittle [16:28] you could use a add/seq/jmp [16:29] i would divide the search into two parts [16:29] 1: search for anything 2: check if its the main part 3: stun 4: wipe [16:30] Join: Neo joined #corewars [16:30] hi Neo [16:31] hi Miz [16:32] damnit, i dont want to give an answer straight out [16:32] I took a quick look at the logs and saw the challenge... :D [16:34] Kalle: lets put it this way: you need only one instruction to differentiate between a clone and the main program, because both the instructions in a clone have something (a number) that main warrior doesnt [16:34] i'm curious... ¿How a simple djn train scores against that warrior? [16:35] so warrior would progress: search for something -> is it clone? return or continue -> stun -> wipe [16:35] Neo: i think better than it should, since the clones (clears) are created with step of only 200 [16:36] Neo: if it had better spacing, it would score badly. [16:38] I see. Looks a very simple opponent. [16:39] i think a simple blur scanner or oneshot would knock it easily [16:40] *knock it down* i mean [16:40] Neo: thats a bit too high solution for this problem [16:47] got something... (i can't resist) :) [16:49] can i post it here? [16:49] you know, since it uses step of 200, it really doesnt matter what you search, if you use small enough pattern, the first thing you find is going to be the spawner anyway [16:49] Neo: if its for school project, i think that would be a little rude [16:50] No it would.. :p [16:50] haha. It is a noob-style proposal :) [16:50] wouldnt* [16:50] The problem I have is to find the core program before his clones kill me [16:51] tip: try a linear scan backwards looking for something. Stun what you find and keep looking 200 locations back. [16:53] yah, that tactic [16:53] if you don't find anything more you have already hit the main code, and then you can start clearing all the core [16:53] can try that in the next warrior. Gonna see how this works out. Atleast im 50%+ [16:55] ok, good luck with it. I'm back to work :) [16:55] Nick Change: Neo changed nick to neo_work [16:57] there, a version that wins 75%, loses 25% [16:59] only real difference is that i changed the scan loop a bit and did a silly trick on the coreclear [17:00] no idea how to change the scan loop so Im just adding a protection on the coreclear and adding a 2nd one [17:03] heres a cheat http://p.cout.se/52 [17:04] however, as that is my code, you might need to do something differently for showing at school [17:04] I just wanna understand it :p [17:06] run it in corewin step-by-step [17:30] MSG: Quit: http://www.mibbit.com ajax IRC Client [18:00] Part: Kalle left #corewars [21:29] Join: OoS joined #corewars [21:29] hi John [21:31] Hi Mizcu [21:32] * OoS still has toothache [21:32] =/ [21:32] I'm sat here with half a mouth of whiskey [21:32] ive been spending half a day tutoring newbies [21:35] :-) [21:45] I can't get into a dentist. Not even the 'emergency' dentist [21:45] I've been trying since Tuesday [21:54] MSG: Ping timeout: 245 seconds [21:56] Join: OoS joined #corewars [22:10] MSG: Ping timeout: 245 seconds [22:12] Join: OoS joined #corewars [23:31] * OoS checks out his web hosting company's 'unlimited' bandwidth offer. [23:31] thats always silly [23:31] jeez, your still awake? [23:32] Yes, still awake [23:33] (im still tutoring zyph here by the side) [23:33] I'm getting about 30Gb traffic on my free host [23:34] Now I'm checking whether the hosting company I use for corewar.co.uk will allow 300Gb on it's unlimited offer.