[00:05] -> zzz [00:24] MSG: Quit: Trillian (http://www.ceruleanstudios.com [05:45] Join: Mannerisk joined #corewars [06:06] Part: Mannerisk left #corewars [15:03] herm. [16:19] Join: Metcalf joined #corewars [16:19] Hi :-) [16:20] h [16:20] Hi Mizcu [16:27] MSG: Ping timeout: 245 seconds [16:28] Join: Metcalf joined #corewars [16:29] I need a hack to make pMARS run on a Nintendo DS [16:30] and you are going to code with a pad? [16:31] Write on the touch screen :-) [16:32] mov.i #((-imp+3*magic)-magic2), }impstep [16:32] try write that [16:47] MSG: Ping timeout: 245 seconds [16:54] hrm. [20:33] wat u hrmin 4 [20:35] its probably subconcious [20:37] Join: Metcalf joined #corewars [20:37] Hmmm... [20:38] one thing i have to say is "darn", Kline's slowscan can get up to 16 triplets, which is about 4 triplets less than what i need [20:39] I'm sure you can add more triples :-) [20:40] how many modifying jmp-variations there are for qscans, or, how many mutations can you get out of single jmp-equiv? [20:42] I'm hrming about my evolver [20:43] Metcalf: let us assume that we are conservative about potential a-fields [20:43] (as in, jmp table as seen in Q^4) [20:50] Can't you have any number of triples, they just get slower? [20:52] well, Kline doesnt mention more than 16 trips in the article [20:52] but you are probably correct, i havent done in-depth analyzation [20:56] thats why i wanted to know the jmp-variants, to calculate the amount of direct jumps and indirect jumps that might be needed [20:57] i wouldnt mind a two-instruction decode if it can be done without jmp-chaining [21:04] I think the jmp chaining is more compact than a two-instruction decode, since the latter would probably need a second table [21:09] MSG: Ping timeout: 245 seconds [21:20] Join: Metcalf joined #corewars [21:22] well, Ivner's Q^2 does 18 triplets with 3-ins table and 3-ins decode [21:23] can an ins be dropped from the decode? does it even do a clean 100-step scan? [21:26] MSG: Ping timeout: 245 seconds