From news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!usc!howland.reston.ans.net!Germany.EU.net!zib-berlin.de!irz401!news.tu-chemnitz.de!uni-erlangen.de!lrz-muenchen.de!news Fri Jul 5 13:56:45 1996 Article: 5564 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!usc!howland.reston.ans.net!Germany.EU.net!zib-berlin.de!irz401!news.tu-chemnitz.de!uni-erlangen.de!lrz-muenchen.de!news From: guenzel@extern.lrz-muenchen.de (Bjoern Guenzel) Newsgroups: rec.games.corewar Subject: spiral launch Date: Thu, 04 Jul 1996 20:01:32 GMT Organization: Tichy Works Lines: 54 Distribution: world Message-ID: <4rh46i$ifm@sparcserver.lrz-muenchen.de> NNTP-Posting-Host: line5d.ppp.lrz-muenchen.de I don't know what kind of spiral launch is John's favourite, but I recently made a small change to the standard continuous binary spiral launch. It was often said that executing like stone/spiral1/stone/spiral2 is an advantage, so I changed the spiral launch from blue candle for running with two processes. Very simple, just leaving out some '}' ... The disadvantage is, that I now have to use 18 (=2*9) processes in the imp, instead of 15 as in blue candle. I haven't really tested how much better the new spiral launch does (or if at all...). Initially my idea was to do it with a vector launch like spl #0 mov.a #vector-j1,j1 spl 1,{j2 .... produce as many processes as you need spl 2 j1 jmp }vector+y,}0 j2 jmp >vector+x,}0 vector dat imp,... ... but I couldn't make it work immediately (perhaps planar knows what's wrong...:), and then I realised that it works shorter and faster with binary launch... Bjoern ;redcode-94 ;name The Impian ;author bjoern guenzel ;strategy continuous spiral launch ;assert CORESIZE==8000 ;release 19.5.96 ;kill test istep equ 2667 start spl 1 ;this is only for demonstration spl #0,<-200 ;put two processes here spl 4,<-1 spl 2,#j1 j1 jmp imp,}0 j2 jmp imp+istep,}0 nop #j2,#j3 ;use for extra attack... j3 jmp imp+2*istep,}0 imp mov.i #istep,*0 end start