[01:05] Part: Mannerisk left #corewars [01:06] Join: Mannerisk joined #corewars [01:06] Part: Mannerisk left #corewars [01:06] Join: Mannerisk joined #corewars [01:22] MSG: Ping timeout: 240 seconds [03:30] Join: Mannerisk joined #corewars [06:33] Join: lllll joined #corewars [06:34] Hello [07:06] evening [07:07] Can I ask you a few things? I'm a newbie at corewars [07:10] I'll do what I can [07:11] You're good at this game? [07:14] If you're busy, it really doesn't matter [07:15] I'm not all that good [07:15] but I'm familiar enough with the commands [07:15] ask away [07:16] Well, I've only been playing for four days, is there any chance you could look at my best warrior so far, and tell me if I've done anything stupidly inefficiant. I'm not asking for you to completely rewrite anything or go over it line by line. It's just I see so many warriors that seem to do what I do using half the size! [07:17] And I get the feeling, I might be doing something dum [07:17] or even "dumb" [07:17] toss it in a pastebin [07:17] if I can't help, someone that checks later would be able to [07:18] A what? [07:18] http://pastebin.com/ [07:18] so a 30-line warrior doesn't fill the chat window [07:19] yeah, I'm just trying to figure out the website you linked [07:20] paste the code in the box, click send, paste the link to it in here [07:20] or if it's a relatively short warrior, you can put it here as well [07:21] http://pastebin.com/m43d6f2f0 [07:22] Is that okay? [07:22] yup [07:26] I'm afraid I'm not good enough to be much help [07:26] Okay, thank you for looking anyway [07:27] I'd check back later [07:27] Thank you [07:27] or check the logs after a few hours [07:27] http://www.koth.org/irc-logs/ [07:27] oh, I didn't know about them. Thanks again! I'm new to IRC too [07:27] no problem [07:34] Oh, it did run okay though for you though? It works here fine, just hope it works for other people too! [07:36] works for me [07:37] Great. Thank you! [09:23] Join: fiveop joined #corewars [09:36] Join: Neogryzor joined #corewars [09:37] hello [09:37] Hello [09:37] hi lllll, i took a look at your warrior [09:38] Ooh! [09:38] So, how bad does it suck? [09:39] * Neogryzor is writting.. [09:40] you can make the scanning loop shorter like this: [09:40] scan: seq start-1,start-22 [09:40] jmp woot [09:40] add.f inc,scan [09:40] djn scan,count [09:40] ... [09:40] inc: spl #49,49 [09:41] The scanning pointers are in the same line, scan, so you can add the step to both at the same time [09:41] by using 'add.f' [09:42] umm [09:43] ohhh [09:43] So, change where i'm scanning in the command it's self, instead of having a variable somewhere? [09:44] and you can use 'inc' as a bomb. spl #x splits to itself always, no matter what X is so still stuns if decremented [09:44] yeah, you can have the variables into your code. Redcode doesn't make difference between code and data [09:44] I don't think I understand that bit [09:45] yeah, I know about variables in code, it's just a bit counter intuitive to me so far [09:45] the spl trick? [09:45] I'm not sure what you mean when you say "and you can use 'inc' as a bomb. spl #x splits to itself always, no matter what X is so still stuns if decremented" [09:46] you put 'spl 0' bombs on your opponent at first, right? [09:46] yes, to slow him down [09:47] if that spl 0 gets decremented, for any reason, it becomes spl -1 and looses eficiency [09:48] ok? [09:48] yes, i get that [09:48] the addressing mode '#' in 'spl #x' makes it to split to itself, no matter what x is [09:49] Oh, I did not know that [09:50] there you go. spl # is invulnerable to decrements, (as a bomb) [09:50] so what did you mean about using inc as a bomb? [09:51] the key is to use both A and B fields in your code when possible [09:52] Yeah, I've seen that used in some other peoples warriors I've looked at. [09:52] also, the mutation of the code in corewars is *very* common. You have to get used to that :-) [09:53] hmm? [09:53] ? [09:53] You mean occasional changes by my enemies to my warrior? [09:54] no, i mean intentional self-changes [09:54] oh, got it [09:56] redcode is a vey powerful language, it can make wonders with a very few instructions [09:58] yeah, I've seen some small self copying warriors, and they do really bizzare things in just a line or two. It boggles my mind [10:00] Thank you for your help though, I'm trying to do the things you said right now [10:00] i can suggest a little change to your attack loop if you want [10:01] Okay! [10:02] morning [10:02] hi Mizcu [10:02] Hiya! [10:04] lllll: someting like [10:04] count: add #24,#0 [10:04] mov inc,}start [10:04] djn -1,count [10:04] sub.a #24,scan [10:04] jmp scan [10:06] What part of my program is that for? [10:07] for when i fill everything with dat's? [10:08] when you drop spl's over your oponent [10:09] Well I do spl's twice. Once during my initial quick scan. Then the 2nd time when I fill everything up with spl's [10:11] i see, loops labeled 'woot' and 'stl'.. [10:11] yes [10:12] i notice 's' and 'empty' are exactly the same :) [10:12] empty changes from a spl for the first wipe, to a dat for the 2nd wipe and all wipes after [10:12] mov trueem,empty [10:13] empty used to be a dat. But i found my wins against silk warrior 1.3 went from 31% to 54% if my first wipe was with spls [10:13] you can remove then 's' and use 'empty' insted [10:14] =\ Are you sure? [10:14] sure! aren't they the same? [10:14] Sorry for being annoying. I've only been at this four days [10:14] *checks again* [10:15] its okay, don't worry. Many begginers can't make a scanner after four days [10:15] Well they start the same, but after one wipe, empty changes from a spl to a dat [10:15] it's how I mop up my enemy after I've slowed him [10:16] You should see my first ever warrior. It's terrible! It took about 10 lines and all it did was copy it's self [10:16] i mean, you can use mov empty,@ erm... i suppouse it is ' Is it? [10:17] well, you can't use two different addresing modes at the same time, it doesn't make sense... [10:18] but without the @, won't it change "Scan" instead of where "Scan" points? [10:18] that reports an error when compiled [10:18] No error here, or i would have fixed it [10:19] !?!?!? [10:19] I use CoreWin. Maybe it doesn't mind it? [10:20] You're right, I changed it and it still works [10:20] i use it to, it reports 32 "Bad Opcode in line ..." errors, lol [10:21] In CoreWin? [10:21] yeah [10:21] 33 actually [10:21] Bizzare [10:21] Want me to reupload it for you? [10:22] @where means to use the B-field of location 'where' as a pointer [10:22] yeah, I think I get you [10:23] so you have to choose one of both [10:26] the indulgence of your Corewin astonishes me :-D [10:27] What do you mean? =O [10:27] I just downloaded it from a link I found a few days back [10:27] We could be on different versions maybe? [10:27] I've got 2.3 [10:28] 2.3 too [10:28] my version doesnt complain, it compiles into < [10:28] (2.1) [10:28] Weird [10:28] Windows version? [10:29] hellish XP [10:30] `\/(o_O)\/` I have no idea what's up then [10:30] hehe, sorry xD [10:30] surprising. How is your compiled code lllll? [10:30] Oh, I checked, it just does < [10:30] So, I changed everthing to < and it works fine [10:31] instead of @< [10:31] oh [10:33] ow, dumb me. I copied the code directly into Corewin, line numbers included :-S [10:33] hehehe [10:34] now it works. I see Corewin ignores first extra addresing modes [10:36] running it i see the scanner leaves big gaps between scanned locations when the core clearing starts [10:36] Well there has to be pretty big gaps so it finds large enemies fast! Or is that not what you meant? [10:38] And i just sped up my initial scan by a third now due to your first tip! [10:39] yeah. It is possible to reduce those gaps as the scanning phase progresses. There is some math in it but, hey, it's your fourth day! you don't need to know that now. Your scanner is okay [10:39] Thanks! [10:39] you are welcome [10:41] gotta go [10:41] bye [10:41] Okay, thank you so much for your help! [10:41] MSG: [10:42] Now i've just got to figure out how to put it into my warrior =O [10:47] Join: impomatic joined #corewars [10:49] Hi :-) [10:50] h3 [10:50] Hello [10:50] Hi Mizcu [10:50] Hi lllll :-) [10:55] how do you swap a and b? [10:56] you can just do mov.x 5,5? [10:57] yes [10:58] but you really shouldnt get into a position where you have to do that [10:58] Oh [10:58] Well, I have [10:58] I'll try to figure out how I can not do it =\ [10:59] it adds an instruction you shouldnt; give me a pastebin of your newest [10:59] oh, I'm just halfway through changing the file I already uploaded. It doesn't work at the moment [10:59] Still want to see it? [11:00] well, it would help to see what the problem is [11:00] http://pastebin.com/m45407876 [11:00] Doesn't work yet. Still working on changing bits [11:01] you can use mov s, {scan [11:01] Oh [11:01] add.a #24, scan [11:01] add.f ? [11:02] remove the mov.x, change add #24, scan into add.a #24, scan, and mov's as told [11:03] Just one of the " mov s, both of those after the removed mov.x [11:04] woot mov.a start,scan [11:04] mov.b start,scan2 [11:04] add.a #24,scan [11:04] woot2 mov s,{scan [11:04] mov s,{scan [11:04] seq scan,scan2 [11:04] jmp woot2 ? [11:05] Seems odd to me [11:06] I think I've borked this up somewhere [11:08] the pasted version seems to work [11:08] mmm [11:09] Not here. But I need to work on it anyway [11:10] been watching this a little bit more, and i dont have any idea why you would want to use the mov.x :} [11:11] Oh, well it's like I said. I was half way through changing things. It's still not really working yet [11:11] but thats ok, when you have enough problems wrapping your head around the main problem, some logic will quickly get lost on the way [11:12] Oh!!! I GET IT NOW!!! [11:12] It is { at both [11:13] will changing the seq to seq.ab compare the a to the b and the b to the a? [11:14] No, it compares the a of the first address to the b of the second [11:14] it will compare a to b [11:14] because it a to b is different then b to a is different too [11:14] it -> if [11:14] umm [11:14] Okay [11:25] following parts at the end of warrior are needless: s, count, empty [11:27] Are they? [11:28] im optimizing the warrior here on the side, but im thinking how to hint how it should be changed instead of showing a fixed version [11:28] Oh, thank you! I like figuring it out. But I like tips too =) [11:30] Why, would your verison kick my versions arse? =D [11:31] You've been REALLY helpful anyway [11:31] All this A vs B register stuff was beyond me till you helped [11:37] http://pastebin.com/da1c900a [11:37] That's not a huge spoiler is it? [11:38] no [11:41] Oh, that's clever [11:41] I think *Confusion* [11:43] I like how you combined the spl and the dat #49,#49 [11:43] Nice trick [11:45] I hope I haven't been stopping you doing important things! It's just you're REALLY helpful [11:45] its saturday, i have none [11:46] Have you had any really successful warriors btw? Just curious [11:48] i have been couple times at koth.org NOP-hill [11:48] ive been trying not to just optimize old warriors, so its been a little slow getting new ideas [11:49] my best so far is at 68/939 at koeningstuhl, and position 12 at dynaHill [11:50] Oooh! nice [11:56] http://pastebin.com/d59775cd3 [11:56] maybe that is a little bit less confusing [11:58] yeah [11:58] Well, I definitely see room for getting it smaller like you show [11:59] Smaller = less damage prone [12:22] Join: Neogryzor joined #corewars [12:23] hi [12:23] Hello [12:23] hi imp, Miz and l^5 [12:23] how is your scanner going? [12:24] There's an annoying bug in it i can't find [12:25] But otherwise good. It's really improving [12:25] Well, i have a few minutes while cooking my meal [12:29] maybe i can help to find the bug [12:30] :D I think I just found it [12:36] Hi Neo :-) [13:34] hi imp [13:57] Part: lllll left #corewars [15:57] * Neogryzor waves [15:57] MSG: [17:01] I like the current nano dynahill status :-) http://www.corewar.info/dhill/i94nano/basho47.htm [20:10] I'm working on a redesign for corewar.co.uk, does this render correctly for you? http://corewar.co.uk/test/ [20:11] Yeah. Although i'd prefer the submenu's titles to be slightly more emphasised [20:13] I can do that, no problem [20:14] I haven't change the visible design yet (apart from the menus being slightly higher) [20:15] The main change is from a float based layout to an relative layout. [21:39] MSG: Quit: humhum