Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!ucsd!rutgers!cbmvax!daveh From: daveh@cbmvax.commodore.com (Dave Haynie) Newsgroups: comp.sys.amiga.advocacy Subject: Re: NeXT-bashing party (hit "n" if you're not interested :-)) Message-ID: <20203@cbmvax.commodore.com> Date: 29 Mar 91 18:30:13 GMT References: <4aQk02L406l501@JUTS.ccc.amdahl.com> <1991Mar17.031448.26855@mintaka.lcs.mit.edu> <15L002.N06Qt01@JUTS.ccc.amdahl.com> <1991Mar26.043648.17656@NCoast.ORG> <2577@ria.ccs.uwo.ca> Reply-To: daveh@cbmvax.commodore.com (Dave Haynie) Distribution: na Organization: Commodore, West Chester, PA Lines: 75 In article <2577@ria.ccs.uwo.ca> ptoper@obelix (Andy Nagy) writes: >In article <4aQk02L406l501@JUTS.ccc.amdahl.com>, >kls30@duts.ccc.amdahl.com (Kent L Shephard) writes: >[stuff deleted] >> Not that easy. More pins on a '040 some of those pins are power pins. "Easy" is a relative term. You don't simply plug a 68040 into the A3000. You must design some bus conversion logic to allow the 68040 to talk to the A3000 bus. You generally need to build a bus cycle converter and a bus sizing mechanism (though the bus sizing mechanism could in theory be a subset of the suggested bus conversion in the Motorola appnotes if you're clever about it). >> The '040 doesn't have the full 68882 built into it. This means that >> either you have to trap the instructions and provide emulation via the OS >> or you can recompile all your applications. The '040 is NOT just a drop >> in replacement for an '030. > How much (functionally) has been left out of the 68040 that the >68882 has? Motorola sez "The floating-point arithmetic instructions supported by the MC68040 are an enhanced subset of the MC68881/MC68882 floating-point coprocessor instructions". The 68040 adds instructions that include the data type they're working on (single/double). It also extends the FPU model with exception support, including the FPIAR (floating-point instruction address register), which points to any floating-point instruction prior to execution, allowing trap handlers to quickly find any unsupported FPU instruction. The missing '882 instructions are mainly the trancendental instructions: [H][A]SIN, [H][A]COS, [H][A]TAN. There are also a couple of manipulation functions, and the constant ROM, no longer built-in. This isn't really that strange, in fact, it's pretty common on RISC systems to emulate some or all of the floating point instructions via traps. SPARC is a good example; it left out the multiply and divide instructions in the first hardware implementations, yet still got good performance. >> I happen to like Intel machines for some applications. What the heck -- some people like Pabst Beer, or Madonna. Doesn't mean either is A Good Thing. >>Engineering software. It's cheaper than any other platform, more >>polished and more abundant. Engineering software is certainly cheaper for MS-DOS than for a typical Workstation. It's also, in many cases, barely functional. MS-DOS does not support enough memory to do much in the way of PCB CAD, for example. There are hacks around this, of course, but to really get that on a PC, you need a '386 and either UNIX or OS/2. And that OS requirement eliminates all three states advantages. Amiga CAD software in this area isn't as mature, but it's already getting as-capable. Pro-Net has its advantages and disadvantages over Schema or OrCAD, and it's only in its second generation. Pro-Board isn't quite as flexible as the OrCAD PCB program, but it's a zillion times faster and less buggy (both Pro-Board and OrCAD I've used are version 2.something). I have only recently acquired BoardMaster, but it seems to seriously kick the butt of the majority of low cost PC PCB applications, and it's dirt cheap. And I have yet to find a bug. On the other hand, there are some things which may never show up on any system other than MS-DOS. There's plenty of things that are so restricted in market, they can only survive on a lowest-common denominator type system, which like it or not is MS-DOS. >> /* Kent L. Shephard : email - kls30@DUTS.ccc.amdahl.com */ >Andy Nagy (ptoper@gaul.csd.uwo.ca) -- Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy "That's me in the corner, that's me in the spotlight" -R.E.M.