Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!tut.cis.ohio-state.edu!ucbvax!hplabs!otter.hpl.hp.com!hpltoad!cdollin!kers From: kers@hplb.hpl.hp.com (Chris Dollin) Newsgroups: comp.arch Subject: Re: Re: skip instructions (on 68040) Message-ID: Date: 9 May 91 09:00:24 GMT References: <1182@opus.NMSU.Edu> <1991May05.174307.8952@iecc.cambridge.ma.us> <7b=C02CO071s01@JUTS.ccc.amdahl.com> Sender: news@hplb.hpl.hp.com (Usenet News Administrator) Distribution: comp Organization: Hewlett-Packard Laboratories, Bristol, UK. Lines: 18 In-Reply-To: rbw00@ccc.amdahl.com's message of 8 May 91 23:29:07 GMT Nntp-Posting-Host: cdollin.hpl.hp.com I seem to remember that the CTL Modular 1 had skip instructions (possibly branch instructions as well). Since I think I've lost my assembler manual, the following may be inaccurate. One of the instructions allowed all sorts of tests on the (few) machine registers; if the test succeeded, the next instruction was skipped. (Instrcutions were 16 bits long.) *However*, there was a way of extending the instruction (I think one of the registers pointed at an extension word), in which case more operations were available, *and* the skip could be 1, 2, or 4 (I think) instructions - a strange hybrid of skip & branch. Skip instructions are a jolly neat way of making arbitrary instructions conditional if you can't make arbitrary instructions conditional (as it were). -- Regards, "If anything anyone locks, they'll find it all ready in stocks." Kers. ----------------------------------------------------------------