From news-rocq.inria.fr!univ-lyon1.fr!ghost.dsi.unimi.it!batcomputer!hookup!news.kei.com!eff!news.umbc.edu!europa.eng.gtefsd.com!howland.reston.ans.net!math.ohio-state.edu!cyber2.cyberstore.ca!nntp.cs.ubc.ca!uw-beaver!cs.washington.edu!jayhan Sat Mar 19 15:19:44 1994 Article: 519 of rec.games.corewar Newsgroups: rec.games.corewar Path: news-rocq.inria.fr!univ-lyon1.fr!ghost.dsi.unimi.it!batcomputer!hookup!news.kei.com!eff!news.umbc.edu!europa.eng.gtefsd.com!howland.reston.ans.net!math.ohio-state.edu!cyber2.cyberstore.ca!nntp.cs.ubc.ca!uw-beaver!cs.washington.edu!jayhan From: jayhan@cs.washington.edu (Jay "Thierry" Han) Subject: Re: Multi-pass core clears In-Reply-To: ojastej@devw751.ice.ncr.doe.ca's message of 17 Mar 1994 13:35:12 PST-8 Message-ID: Lines: 57 Sender: news@beaver.cs.washington.edu (USENET News System) Organization: Computer Science & Engineering, U. of Washington References: Date: 18 Mar 1994 11:59:21 GMT Okay, so James as uncovered his Silly's :-) I should also take the plunge. At the end of this article, I put the source for my Sunburst 33, which plays around 10th on the 94x Hill. The Sunburst series (didn't go very far on the 88 Hill) is basically a stone with a multi-pass coreclear. I experimented with most of the things James talks about, and decided that a SPL0 pass followed by ad libitum DAT core-clear was the most efficient. Now I have second thoughts about it, because I think that SPL0 then DAT then perfect gate is better (Those imp-spirals can rise from the dead, you know). Anyway, this warrior was originally written in '88, so it doesn't do forward core-clear. The stone is the "confuse/bomb" type, with constants carefully tailored to cover 25% of the core with either a bomb or a B-decrement. No spot larger than 6 consecutive instructions is left out, so any apponent bigger than 6 instructions is touched, and most opponents twice or more. Notice a few nitbits. At the end of the stone run, p1 points to 1 and p2 points to 2. Thus, it overwrites "mutate" with "ptr". Notice the DJN that triggers the change from a SPL0 coreclear to an infinite DAT coreclear, by redirecting the pointer at "mutate". Also, the advantage of having a core-clear sequence of "SPL 0" - "MOV" - "JMP -1" is that it accumulates processes, which means a tie againts an imp-spiral. I was really happy when I was able to bring down the program's size down to 8 lines. A stone must be small enough to go undetected by a scanner, its target of choice. I'll try incorporating James' idea of forward-core-clear. ------------------------------------------------------------------------------- ;redcode-94x ;name Sunburst 33 ;author Jay Han ;kill Sunburst ;p1 must hit 2 and p2 hit 1 p1 equ 27790 s1 equ 68 p2 equ 6914 s2 equ 55423 loop add inc, confuse confuse mov =- e-mail: jayhan@cs.washington.edu Dept. of Computer Science and Engineering, FR-35. Fax: (206)543-2969 U. of Washington. Seattle, WA 98195, USA. [Sieg 230] (206)685-1224 Personal: 4820 Burke Ave N. Seattle, WA 98103, USA. (206)547-8559