Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!uunet!cbmvax!daveh From: daveh@cbmvax.UUCP (Dave Haynie) Newsgroups: comp.sys.mac Subject: Re: 16Mhz why in the world.... Message-ID: <6461@cbmvax.UUCP> Date: 30 Mar 89 19:22:37 GMT References: <1455@ccnysci.UUCP> Organization: Commodore Technology, West Chester, PA Lines: 75 in article <1455@ccnysci.UUCP>, alexis@ccnysci.UUCP (Alexis Rosen) says: > Enough frothing about the mouth. The other answer only applies to the CPUs, > but it does excuse Apple for the slow processor speeds (though not anything > else). It turns out that Motorola is having a bitch of a time getting the > '20 and '30 out in quantity at high speeds. It appears that they can't > supply enough to Apple to meet the demand. I think there's a simple explanation. The Apple Marketroids wanted an '030 machine, because the "68030" is hot, and "68020" is luke warm. And they wanted such a machine last fall. So Apple took an existing Mac II, dropped an '030 in it, and . The IIx and IIcx really aren't designed to be 68030 machines; they're already running main memory at about 1/2 the speed their 15.8MHz 68030 could actually use, a much faster processor in the same system might not be considered "worth it" to Apple. But IIx and IIcx solved the immediate need for a 68030 machine, and could let them come out with a real one, say at maybe 25MHz or 33MHz, in the future, including external cache or main memory that's at least reasonably matched to the speed of the processor. And they get to charge much more for such a machine, and/or lower the costs of the current line, even if the actualy cost of such a machine isn't all that much more. > Now Motorola can barely keep up with Intel- yes, they're shipping a 33MHz >'030, but in what quantity? And what will it be like in a year, with 40 MHz >'486s? Will the '040 be able to keep up? Motorola says so, but they can't >even get the 030 out in quantity! It doesn't look like anyone else is having supply problems with Motorola. The only problem I've run into is that 25MHz 68881s don't yield too well, but you can get all the 25MHz 68882s you like if you're willing to spend a little bit more (makes me wonder about how much trouble they've gone to making 25MHz 68881s yield better). Considering that a full speed '030 is considerably faster than a full speed '386, and that both Moto and Intel are claiming around 2x improvements for their respective new chips, I claim the gap, if there is one, is widening (Moto actually claimed at least 2x increase for the '040 integer unit, about 10x-20x increase in the floating point unit over an '030+'882 system). > The point is not that I want Macs to be superior to IBMs in raw speed. I do. > And they're not, thanks to lazyness on Apple's part. The point IS that the > architecture we're bound to may, in the end, be a disaster for us. Just two > short years ago, things looked very different. You might wonder if part of the problem isn't laziness on Apple's part. Look at IBM these days. They've felt the pressure of cloners like Compaq and HP, and they're actually starting to build respectable machines (respectable in their class; I personally don't like '386 systems). EVERY high end '386 machine these days has a decent sized external static cache, most support a high speed Weitek floating point unit as well as the Intel FPU. The reason for this improvement in the hardware is direct competition. No one makes Mac clones, so Apple doesn't have any primary competition like the cloners do, only secondary competition with other computer vendors, none of whom are selling the same thing Apple's selling. As long as Apple believes you'd rather have a Mac than a PClone, they don't have to keep up with the technology to the same degree that cloners will. Considering how bad the PClone world is with software these days, Apple may be able to take their time. Macs have never really been known as high speed systems, they've just been winning on all kinds of software and user issues. IBM used to think it was pretty much secure just by being IBM; these days they're main efforts are in high speed and new software. > (BTW, watch for the announcement on the '040 later this week. I hear they > say they'll have quantity 25MHz parts late this year, turning in about > **10 MIPS**. We'll see...) Last figure I was 12.5 MIPS, 2-3.5 MFLOPs. Tasty. > Alexis Rosen -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy Amiga -- It's not just a job, it's an obsession