Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucsd!sdd.hp.com!uakari.primate.wisc.edu!aplcen!jhunix!jheddy From: jheddy@jhunix.HCF.JHU.EDU (Jared Brennan) Newsgroups: comp.sys.atari.st Subject: Re: Atari TT 030 Launched! Summary: 33% increase: it's the line a traps, isn't it? Keywords: Atari TT Message-ID: <5503@jhunix.HCF.JHU.EDU> Date: 11 Jun 90 04:39:40 GMT References: <1990Jun5.143231.4977@watserv1.waterloo.edu> <13266@wpi.wpi.edu> <81214@tut.cis.ohio-state.edu> <1990Jun6.044350.20403@cbnewsh.att.com> Reply-To: jheddy@jhunix.UUCP (Jared Brennan) Distribution: usa Organization: The Johns Hopkins University - HCF Lines: 18 The 33% speed increase in Atari ST programs run on the TT has aroused quite a bit of interest on the net recently, because this is rather minimal compared to the speed increase expected given 2X clock rate, 2X data bus size, caching, etc. However, I believe that the line A traps are real instructions on the 68030, while they are invalid on the 68000, which allows them to be used for ST graphics primitives. So, contrarily enough, the line A traps are "illegal" instructions on the 68030 TT, while they work fine on the 68000 ST. The result being that you have to go through some gymnastics to run ST programs on the TT. Hence, the limited speed increase. I think this was discussed on the net some time ago. I wish somebody from Atari would toss in their own $0.02, because I'm really uncertain about this information. -- . . . Yet another shallow and badly thought-out opinion from: Jared Brennan BITNET: JHEDDY@JHUNIX,JHUVMS