[00:28] mfing got it! [00:33] http://www.nomorepasting.com/getpaste.php?pasteid=20865 [00:35] the hacks hurt my head [00:35] but it works [00:35] i can't figure out why the thing with the s variable above doesn't work [00:35] so instead i counted the instructions needed and made it only loop that many times [00:53] hee. 63 point imps [00:54] Nimbus? [00:54] no [00:54] just playing with the thing i wrote [00:55] i wonder if anyone made a 2667 point imp [00:55] lol. [00:55] since you're back, maybe you can explain the problem with s above [00:58] as said, i have no interest in how the preprocessor of pmars/corewin works [01:00] oh, right [01:00] still, i hoped maybe it was something you knew about [01:00] :P [01:00] i like this way much less [01:00] it's doubly wasteful [01:30] anyway, you saw it right? :) [01:34] ? [01:34] http://www.nomorepasting.com/getpaste.php?pasteid=20865 [01:34] yeah [01:35] it's not very cool? :( [01:35] id rather debug compiled qscan-code [01:36] haha [01:36] it doesn't need debugging, it works [01:36] just set values in procs and arms [03:07] wtf it always kills the same imp no matter where i put it o_O [03:09] what kills? [03:09] stone killing my imp [03:09] :( [03:10] i was reading it wrong [03:10] i thought that the imp survived the other parts of the clear [03:10] but i was not correct [03:10] i added more procs and it lives now [03:12] weird, mirc is eating cpu [03:12] corewin is going slow because of it [03:18] ow, it owns my highest scorer so far [03:18] lol [03:18] let's see how it does on the hill [03:19] #4 [03:19] huh [03:19] i expected it to suck [03:19] :> [03:31] well you were right, it did good against the clears (slear disappeared while i was working, boo) [03:31] and the papers, well *shrug* [03:32] i'll take it home and see if optimax gives me any helpful numbers [06:10] Join: Davse_Bam joined #corewars [06:47] MSG: Ping timeout: 240 seconds [07:40] MSG: Quit: leaving [07:43] Join: pak21 joined #corewars [09:41] Join: fiveop joined #corewars [10:09] Join: sh0ne joined #corewars [10:10] hi [10:48] hi :-) [10:49] MSG: Quit: humhum [11:25] Join: Altair joined #corewars [11:25] hi [11:43] hi alexander [11:43] >_> [11:43] or Altair [11:48] Join: Mizcu joined #corewars [11:49] im not a happy camper about this ISP [12:38] Join: fiveop joined #corewars [13:11] MSG: Quit: Leaving [14:10] Join: pie joined #corewars [14:13] MSG: Client Quit [14:47] MSG: Quit: Read error: Connection reset by peer [16:08] Join: bvowk joined #corewars [16:08] hrm. [16:27] Join: sh0ne joined #corewars [16:28] Hi [16:29] how's it going shone? [16:33] bvowk: It's OK, a little pressure from job, a little from girlfriend...well, it's ok i guess.. [16:37] Anything exciting on the corewars front? [17:03] nil [18:02] if I use djn.f bla1,bla2 have both values of "bla2" have to be 0 ? [18:02] y [18:03] is there an option to dekrement both, but jumps if only one is zero? [18:03] no [18:19] hm i need another factor in my stone/imp [18:19] it seems so ugly to just paste the process generating code twice [18:20] but maybe that's what i should do, then i can run it through optimax and see what balances best [18:20] *sigh* [18:21] ? [18:23] you are overcomplicating the issue [18:23] nah i just want to see what happens [18:23] what am i overcomplicating? [18:24] none of my stone/imps have had their processes optimaxed [18:24] or optimaxing varying instructions [18:25] but its so fun to say! :P [18:25] it's just an easy way to test a variety of things [18:25] i don't know what i "want" [18:25] only things actually have been the boot-distances (in case of Static, it was only one number, since second was relative to first) [18:25] do i start the stone first or after launchin the imp? do i feed it more than one process to start? do i use a spl 0? etc [18:26] so i'm interested in the results from different process balances [18:26] compare the values by hand-benchmarking [18:26] if you have a better way please tell me :P [18:27] i dont have a better way, i only have a way which requires you to dive instead of just floating [18:28] by which you mean doing it the long way? :P [18:28] there is no need to check all the possibilities [18:28] i don't want to check all the possibilities [18:28] * nescience shrugs [18:28] a cat is fine too! i can tweak things by hand [18:28] take, say, 6 variations, find the one that scores best, then try some small variations in the allocation [18:28] whenever i have the time to do it [18:29] also that means i have to figure out how to benchmark things lol [18:29] what OS did you use? [18:29] windows [18:29] but i could emulate some linux i suppose [18:30] corewin has a benchmark-option [18:30] load up your warrior plus a benchmark, choose benchmark: yourwarrior, start a run [18:30] there didn't seem to be a benchmark-file for corewin [18:30] in the fsh benchmarks etc [18:31] is it just a list of paths? [18:31] fsh is too heavy [18:31] when you have ~100 warriors, it blurs the point where you can see the actual changes [18:36] can't, you mean? [18:36] what would you recommend? [18:37] there arent that many ready benchmarks people have built [18:38] wilkies (ancient) wilmoo, wilbez (i dislike because they have p-spacers with, distorts scoring), wilfiz (on the challenging side), wilmiz (havent officially released it yet, and "average" benchmark) [18:40] and -> an [18:40] k [18:41] wiluorwontu! [18:56] what do you think about this stone: [18:56] http://www.nomorepasting.com/getpaste.php?pasteid=20886 [18:58] using the spl to store steps isnt really common, but i wont intervene on that [18:59] isn't common or isn't a good idea? ;P [18:59] hint: add.x inc, @-1 , the last bomb hits atk+1 [19:00] it doesn't bomb itself, it bombs the increment first [19:00] or else i'm off by one step perhaps [19:00] spl mov mov add djn when it enters the clear [19:00] that was a suggestion, not an analysis [19:00] oh [19:01] that overwriting of inc isnt bad either [19:01] i'm a little confused [19:01] you WANT to bomb atk+1 ? [19:01] thats one possibility [19:02] or do you mean if it happens [19:02] sooner or later you are going to hit yourself, might as well make the hit useful [19:02] but overwriting the inc was something my tired mind didnt notice [19:02] does pretty much the same [19:03] aha [19:03] i can't parse what that @-1 would do though lol [19:04] if it got bombed, it'd add inc to itself [19:04] er i mean, if atk+1 did [19:04] oh i get it [19:04] spl, mov, dat [19:04] the dat increases the mov for a clear [19:04] ? [19:04] yes [19:05] well it wouldn't hurt anything either way [19:05] if the poitner went awry there's a slight chance of it doing something useful [19:06] im thinking in terms of Devilish-style stone, which is different enough from that not to be directly comparable [19:07] i have to look up every warrior you reference in that fashion ;) [19:09] i was pondering how to make it do better against papers [19:10] wouldn't really hurt anything to use an spl-dat clear or something if i could hack it right? [19:10] good luck getting a stone to turn into a spl/dat -clear [19:10] Join: CoreOld joined #corewars [19:11] lol [19:11] just sayin' [19:11] (Winter werewolf does bomber -> SD) [19:11] since it was throwing around bombs with 1s and incs, i thought i might be able to get something to change the bomb poiter [19:11] short, wide bombing run -> SD-clear + imps is a possible combination [19:14] is a bomber not a synonym for a stone? [19:14] it used to be [19:15] now the consensus is that dat-bombs -> stone custom bombs -> bomber [19:15] huh [19:15] ok [19:18] me goes zzz [19:19] 'night [20:06] MSG: [20:08] Join: nescience joined #corewars [20:09] woops [20:09] forgot i had no way to get reopped [21:16] nescience, are you the one who wrote the Imp Alpha Engine on SAL beginner hill`? [21:23] MSG: Quit: Leaving [21:30] Join: sh0ne joined #corewars [21:31] MSG: Quit: Trillian (http://www.ceruleanstudios.com [21:48] yes [21:49] i lol'd because it launched at #3 and now as more entries came in it worked its way up [21:49] i improved it by 15 points against the benchmark i'm looking at right now though [21:49] was about to resubmit [21:51] and the name isnt fake? you are just using imp spirals to kill the opponent o.o? [21:51] oh sorry [21:51] i'm not in the habit of adding strategy lines yet [21:51] i added a stone to the hill version [21:51] it's just a stone imp [21:51] the 'engine' part is the stuff i wrote so i can just be like [21:52] arms equ X [21:52] procs equ Y [21:52] and it makes the spiral for me [21:52] k. I really was wondering, how an imp could kill THAT good^^ [21:52] nice. [22:04] well now you know the answer :> [22:04] it throws wrocks [22:04] rocks [22:09] MSG: [22:30] MSG: Quit: Raided by Secret Service [22:46] i scored a 120 and now i don't remember how lol [22:57] MSG: Quit: humhum [23:06] Join: Rashnok joined #corewars [23:07] MSG: Client Quit [23:47] Join: Rashnok joined #corewars