From news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!cs.utexas.edu!swrinde.nde.swri.edu!nntp.primenet.com!enews.sgi.com!insync!news.azstarnet.com!news Mon Sep 16 10:25:32 1996 Article: 5944 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!cs.utexas.edu!swrinde.nde.swri.edu!nntp.primenet.com!enews.sgi.com!insync!news.azstarnet.com!news From: Franz Newsgroups: rec.games.corewar Subject: Re: a new opcode ideae Date: Sat, 14 Sep 1996 15:24:22 -0700 Organization: StarNet Lines: 125 Message-ID: <323B3096.7D45FB73@azstarnet.com> References: <960914022812_102741.2022_GHT101-2@CompuServe.COM> NNTP-Posting-Host: usr7ip55.azstarnet.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; I; Linux 1.2.13 i586) > It would be nice to make the program execution step instruction less powerful, > like Set Execution Step(SES?) x, and that would be all it does. The imp > possibilities look too powerful. (How to kill an optima number imp? A gate > won't do it...) What if it set the execution step for all processes though? > Make it harder to use fancy imps and normal code at the same time. > > Justin > > P.S. I think it should not be called xxx increment, because its not being > incremented anymore, its really a step. (Like BASIC: FOR I=2 TO 1000 STEP 2) well the ses instruction would have one disadvantage you could still make an optiman number imp, just boot it and then set the execution step .. just two more instructions for startup ... but it would kill the possibility for a chaser .. well maybe not how about this piece of code MOV -3,<-3 start SPL 2 SES 0 MOV -3,-1 this would be the single instruction core clear and similarly could be done the corechaser ... i still haven't thought of it ... but a 75%c bomber is possible like tornado without the jumps and with JSI 1,#0 type instruction booting ... anywayz ... there woudl be a better way to make the game fairer ... add anotehr limitting factor ... the maximum increment ... it would be set on the command line and the program could not increase it's increment/step .. beyond that ... if it was somehting like 5 allowing ranges for -5 to 5, including 0 ... that would forbid the optima num imp, and would make most bombers distributed over 2-5 step width imune the the imps .... at CORESIZE==8000 an imp could do only a 3 step to cover all positions and in a stepped warrior that's easy to kill with a gate ... just count mod3 number from the possible targets .... and make yer warrior mod 3 distributed ... you can test hpw that works with the current test source or i can put this thing in to make you feel better :) though i think there still is a possibility of killing an optima IMP ... it still has not got a chance againt a paper ... well some chance but i think about as much chance as an optima numebr bomber ... these are the results i got for thsi type of imp with a few warriors at, 200 rounds these are normal warriors Optima IMP by Anonymous scores 134 Flash Paper3.7 by Matt Hastings scores 398 Results: 22 110 68 Optima IMP by Anonymous scores 440 Rave by Stefan Strack scores 155 Results: 145 50 5 Optima IMP by Anonymous scores 298 Simpap by Franz scores 256 Results: 84 70 46 Optima IMP by Anonymous scores 233 Greed by Franz scores 308 Results: 58 83 59 Optima IMP by Anonymous scores 24 IMP Spiral by Anonymous scores 555 Results: 1 178 21 This is the prototype mod 2 chaser using JSI instructions ... I jsut wrote it in the last 3 minutes, nothing fancy ... if your warrior is mod 2 distributed adding these lines of code will kill imps Chaser by Anonymous scores 472 Optima IMP by Anonymous scores 64 Results: 136 0 64 here's the chaser code ... quick and dirty ... i bet it migth be improved ;redcode ;name Chaser ;assert CORESIZE==8000 SPL addme JSI 1,#0 MOV {data,