Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!usc!rutgers!bagate!cbmvax!daveh From: daveh@cbmvax.commodore.com (Dave Haynie) Newsgroups: comp.sys.amiga Subject: Re: Sorry Guys, There is NO WAY! Message-ID: <16432@cbmvax.commodore.com> Date: 11 Dec 90 00:40:10 GMT References: <9012032130.AA114840@vttcf.cc.vt.edu> <21968@well.sf.ca.us> <1990Dec5.194205.16892@sbcs.sunysb.edu> Reply-To: daveh@cbmvax.commodore.com (Dave Haynie) Organization: Commodore, West Chester, PA Lines: 56 In article <1990Dec5.194205.16892@sbcs.sunysb.edu> dtiberio@libws3.ic.sunysb.edu (David Tiberio) writes: >IF I REMEMBER CORRECTLY, THE AMIGA 3000 HAS A FULL SOCKET >When the 68040 is out, it should pop right in to where the 68030 was. The A3000 does not have a 68040 socket onboard. What it does have is a 32 bit Coprocessor Slot, and consideration in the board logic given to what a 68040 might want to do that a 68030 can't do. For instance, while the 68030 only supports burst cache line fetches, the A3000's onboard memory system supports cache line dumps as well, as does the '040. At the board designer's option, the 68040 can run synchronous with the motherboard at 16MHz or 25MHz, or it can run asynchronous at any clock speed. The motherboard's 68030 can be used or turned off, depending on how the coprocessor board and system software choose to treat things. This is the same basic approach we took with the A2000's Coprocessor slot, only we knew considerably more about the 68040 than we did about the 68030 at A2000 design time, and of course the '040 and '030 are much more bus compatible to start with than the '030 and 68000. A coprocessor board does, however, have to be designed by someone who knows what they're doing. There is a bit of interface logic needed for any coprocessor board, there are quite a few hardware options to choose from, the issue of on-board memory or cache, etc. >When this happens, the Amiga 3000 will run at 15-20 MIPS, will not require >the slow 68882, etc. Don't compare a 68030 with a 68040! Again, a 68040 system MAY run 15-20 MIPS, it may not. What Motorola says is what their marketing department considers possible, based on input from the technical people. And in my experience, they've been correct about what's possible in the past. If the 68040 systems that actually get built have as many wait states as the 68030 systems that got built, you won't see this performance in real life. >The Amiga 3000 should have a socketed 68030, with extra pins for "future >expansion". I take this to mean that they planned it to work with the 68040 >and the 68030. No, the 68030 isn't socketed, it's surface mounted and stays in place. You don't go plugging a 68040 into a 68030 socket. If you have to kludge it, you might build a 68040 board that could live in a 68030 socket, but that's going at it backwards -- the capability to live a nicely in the system as possible should be provided from the beginning. An expansion card connector like the A3000's 200 pin slot is designed to handle a plug-in card, the 68030's socket is not. >The designers are sure to know. We better! >David Tiberio SUNY Stony Brook 2-3605 AMIGA Toto Productions DDD Men -- Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy "I can't drive 55" -Sammy Hagar