[03:30] MSG: Quit: Lost terminal [07:34] typical, just typical [07:35] its supposed to be an examination week, but yet the school network is slower than an 8am in Friday [08:53] Join: Waknuk joined #corewars [09:02] MSG: Remote host closed the connection [12:14] Join: fiveop joined #corewars [16:27] Part: ares left #corewars [17:21] morning. [17:22] ah man [17:22] got pushed back down to #25 [17:25] hrm. [17:26] got my score went up from 25 to 64. [17:27] pompis is a scanner, throws jmps. does a one pass scan, then splits to do another pass and a core clear. [17:42] Join: John joined #corewars [17:42] Hi :-) [17:44] greetings [17:44] Hi Jon_ [17:47] Bvowk: did you say your script for running a 150000 warrior round robin is Perl or Python? [17:48] I have something I'd like to run! [17:49] Does anyone have problems running nMars? [17:51] haven't tried it [17:51] does it run under mono? or just .net on windows? [17:52] That's doesn't mean anything to me! :-) [17:52] ah [17:52] mono is a .net implementation for linux. [17:52] Oh, I see. Not really sure what .net is though. [17:53] it's microsofts java ripoff :) [17:54] Do people still use MS programming languages? [17:54] yep. [18:03] back [18:05] Hi Mizcu [18:07] hrm [18:08] John: python [18:08] bvowk: I'll have to write one in Perl [18:09] why dont you do it with basi--- oh, you already did that [18:12] Because I want to run about 30000 warriors round robin [18:12] And the basic program can only copy with 2000 or so [18:12] copy EQU cope [18:13] met: personally, I'd use python, because you could use pycorewars.. and it'll be significantly faster [18:13] write one in assembler :) [18:16] I think the only place that would pay off is if you could figure out a way to run multiple rounds concurrently with wide SSE instructions [18:17] I mostly wrote one in assembler. I just didn't like the way it had to call pmars :-( [18:19] then use exhaust [18:20] challenges are supposed to be fun [18:29] * jon_ looks up SSE on wikipedia [18:36] I'm sure hotmail is gradually removing features [18:37] how come? you trust the odds, or they are using you as their first target? [18:40] Stuff I used to use can't be done anymore :-( [18:41] Used to be able to click on the sender or subject to sort messages, can't do that now. Used to be able to search for a phrase in the message body, now it's limited to searching the subject [18:41] what browser you use? [18:45] dont expect hotmail to work on a good browser [18:45] even though it doesnt whine about Opera, which is a miracle [18:46] (but once i spoofed my browser identification to bypass a download limit, forgot it on, and hotmail started blathering) [18:48] I'm using Firefox [18:48] problem obvious, end of case [18:48] :-/ [18:48] I'll try IE then [19:01] Join: Fluffy joined #corewars [19:02] :) [19:03] Hi Fluffy [19:03] * bvowk mocks fluffy needlessly. [19:04] * Fluffy taunts bvowky for no good reason [19:04] Hi John! [19:04] Fluffy: i'm still on the bottom of the beginners hill, but my score is up to 80 now :D [19:04] jon_: Nice! What kind of warrior is it? [19:05] bvowk: Look at the nano-hill. There's nothing more to say! [19:05] scanner that changes to a coreclear after 1 pass. [19:05] How long is that pass? [19:05] scans 50 locations [19:05] 50 locations are too few! [19:06] Try some 1000 locations [19:06] not mu-- darn floff [19:06] hmm [19:06] before switching to the clear [19:06] When you have working warrior it often pays to play with its "values" [19:06] (like the number of scans or bombs, ...) [19:06] Mizcu: yup :) [19:07] i wanted to scan quickly as i figured if i'm sitting in that loop too long, im a sitting duck. [19:07] Yes, that's what all scanners are [19:07] but usually it is a good idea to scan a little bit more [19:07] hmm [19:07] ok [19:08] a normal qscan does only check about 30 ... 50 locations, a scanner should do more [19:08] ill try 500 [19:08] yes, try different values :) [19:08] fluffy: so? you put your old warrior back on top, call me when the flying sink goes straight to the top [19:09] Fluffy: when i find something i drop a dat bomb, i tried dropping a nasty jmp before. which do you think would be more effective? [19:09] bvowk: I could do that, but that's not what I want to do [19:09] or spl 0,0.. [19:09] etc. [19:10] dropping dat's works only on nano [19:10] jmp works only on LP [19:10] I hope that it is an "SPL # 0, whatever" and not a "SPL $ 0" [19:11] :) [19:11] you can try throwing a "spl # 0 ; jmp -1" [19:11] silk-papers copy too fast to be harmed by few dat's or to care about jmp's [19:12] ok [19:12] i just submitted pompi-fluff [19:12] scans 500 locations, drops spl #0 bombs [19:13] Good luck [19:14] of course, throwing a zillion dat's can kill a paper, but its not as reliable as a good scanner [19:15] lol [19:15] scored 56. [19:16] jon_: hmm ... maybe a bug or you hit yourself with that split bomb [19:17] hmm [19:17] i will continue this outside of business hours. [19:32] From the next nano Hall of Fame update I'll just record warriors over age 200 [19:32] > I'll put a copy of the current HoF in the old directory http://corewar.co.uk/old/ [19:35] That's probably a good idea [19:36] Average age is higher on a hill with 50 warriors anyway [19:39] Right, got to go [19:39] * Fluffy waves [19:39] I've made a list of Corewar stuff I'm going to do each month, and I've only just started this month's :-( [19:39] * John waves [19:39] NanoWarrior? [19:40] Yes nw is on the list for this month [19:40] :) [19:40] MSG: Quit: mov.i #1,1 [19:44] given the number of challenges to nano.. a 500 is almost as good as a 100 on other hills [19:45] hmm [19:47] you don't agree? [19:48] [3jane:~/work-tmp/jnsl$] find clean_save -name \*.rc -type f |wc -l [19:48] 707373 [19:48] 700k nano warriors.. [19:48] scary. [19:48] the average age of a warrior on a 25-slot-hill is 25; the average age of a warrior on a 50-slot-hill is 50. If we use an age-limit of 100 for the 25-slot-hill, then an age-limit of 200 is the right one for a 50-slot-hill [19:49] I don't think the age scales linearily with the number of slots. [19:49] it does [19:50] not really, because the hill changes more slowly. [19:50] so superior warriors have their niches live longer than smaller hills [19:51] simply because new challengers have to perform better against a larger crop, so most only squeak in the bottom edge [19:51] that doesn't change the average age of a pushed off warrior [19:51] no, but it changes the distribution [19:51] and the hall of fame isn't just for your average warriors now is it? [19:52] hmm ... let's simply raise the limit to 1000, ok? [19:52] ;) [19:52] heh [19:52] damn ... then you wouldn't have any warriors in the nanoHoF [19:53] I really really didn't see that before ;) [19:54] so you're all for the idea then? ;) [19:55] http://sal.math.ualberta.ca/hill.php?key=nano [19:55] No, that would be a really short HoF [19:55] holy crap! that's unexpected! [19:55] huh? [19:55] that warrior is utterly tuned for the wrong hill! [19:56] I was expecting it to kick ass, not get its ass kicked.. [19:56] yes, so it seems [19:56] what was its score against your current benchmark [19:56] ? [19:56] this seems to indicate the skimming technique is flawed. [19:56] it was wildly ahead of the pack by like 5 points [19:56] what kind of skimming technique? [19:57] 161 vs 156 for the next warrior [19:57] for keeping a balanced hill while reducing the number of warriors in my valhalla [19:58] the full valhalla is balanced pretty closely to nano, but the valhalla thats been trimmed to about 1/5th the size is skewed badly it looks like :( [20:00] I use more or less all published warriors, which made it into the nano-HoF plus some of my own warriors [20:00] that seems to work pretty well so far [20:10] grmpf ... That's already the second time that an OpenOffice compile failed :( [20:14] compiling OO is an adventure. [20:14] Yes, but it usually works. [20:15] At the least the normal LFS (Linux from Scratch) howtos work. [20:16] where's the HoF? [20:16] corewar.co.uk [20:16] http://www.corewar.co.uk/logs.htm [20:16] Fixed on TM_20070312_RC1 [20:16] [20:16] woops [20:18] no, just the wrong window [20:18] middleclick [20:47] * Fluffy waves [20:47] MSG: Quit: fluffy.i < 1, # 42