hes@ecsvax.UUCP (Henry Schaffer) (02/12/86)
<>I was looking through the IBM January 21, 1986 Announcement
Letters for the RT PC, and was particularly interested in the
cpu. "The system processor has a 32 bit, reduced intstruction
set computer architecture, developed by IBM on a single chip
^^^^^^^^^^^^^^^^
using 2-micron FET technology. It has sixteen 32-bit general
purpose registers and uses 32-bit addresses and data paths. The
microprocessor is controlled by 118 simple 2- and 4-byte
instructions. An IBM-developed advanced memory management chip
^^^^^^^^^^^^^
provides virtual memory address translation functions and
memory control. It provides a 40-bit (1 terabyte) virtual
address." (^^^^emphasis added)
I didn't see anything in the announcement about availability
of these chips by themselves. If these chips are not
available, it might be very difficult (impossible?) for
competitors to build directly competitive machines - they would
be limited to supplying add-on boards. This is quite a
difference from the present situation with the 8088 and 80286
based micros, because of the use of an available cpu chip.
Of course in the good old days every computer had a
proprietary instruction set - have we come to expect a
different situation?
--henry schaffermark@dmcnh.UUCP (Mark Roddy) (02/14/86)
> > I didn't see anything in the announcement about availability > of these chips by themselves. If these chips are not > available, it might be very difficult (impossible?) for > competitors to build directly competitive machines - they would > be limited to supplying add-on boards. This is quite a > difference from the present situation with the 8088 and 80286 > based micros, because of the use of an available cpu chip. > > Of course in the good old days every computer had a > proprietary instruction set - have we come to expect a > different situation? > And in the good old days, the plug compatible manufacturers made quit a bit of money underselling IBM. Proprietary chip sets raise the ante to enter the game, but they don't make it impossible to be competitive.
rb@ccivax.UUCP (rex ballard) (02/19/86)
> Previous article mentioned that IBM developed RISC chip is not available > separately. If IBM is up to their usual tricks, this (Proprietary RISC chip) is what they hope will be come the "New Standard" in "PC's"! Certainly effective against the "Clone market"! It's amazing how many DP administrators have gone along with this too! The one thing different this time is that PC purchase decisions are/were made higher up. Deja Vu? (Inside jokes for those who have supported or bought IBM in the past) The latest "Plug Compatible" - the 360!!! The latest "Plug Compatible" - the 370!!! Remember EDX? Upgraded your Series 1 yet? Which 370? What bundling? Upgrade or Replace? Is it (PC-RT this week) IBM compatible? IBM will change no rules before it's time! And now for this years replacement for last years product... Who's gonna tell the Pres. the IBM we bought last month is obsolete! Seriously folks, when IBM says "We won't change the rules", that fits in the same catagory as "The check is in the mail" or "I won't !@# in your face". Actually, to be fair, other manufacturers respond the same way. Apple, Commodore, and Atari upgraded their machines in response to IBM. This is just your normal competitive response. When IBM came out with the PC, the value of a Series 1 deteriorated rapidly. But the PC was an effective response against the CP/M desk-tops dominating the market at the time. MS-DOS may seem a bit "plain and vanilla" right now, but the CP/M boxes were pretty much "Text Only" with little support or standardization for even simple cursor positioning. The boxes with graphics relied on "Basic in Rom" operating systems so proprietary they weren't even source code compatible. Disk formats ranged from 35 track, single density, single sided to 96 track, quad density, double sided with 1 16KB sector per track. IBM eventually went with the quad density floppy, but the 360K was reliable, provided reasonable storage for "floppy only" systems. The graphics hardware, mountable drivers, interceptable interrupts, and "open architecture", opened the doors to hardware and software technology that was only science fiction as little as 2 years prior to its release. MS-DOS (PC-DOS) boxed them in, competitors came out with better or lower cost equivelant products and IBM is responding in their usual way. IBM needed a machine that would offer multi-tasking, VDI real-time graphics, high speed, mass storage, and a large linear address space. I would assume that the 68000 was considered too 'generic', and lacked the support (Memory management, floating point, caching, video support circuitry, etc) to be a candidate for a "dramatic new design". A RISC chip also makes bus sharing with very high resolution displays or very high speed DMA peripherals and co-processors more practical as well. Fortunately for IBM, the 68K boxes have not adopted any real "standards" yet. This leaves the market open for IBM to repeat the PC success again. If the 68K producers get things together and adopt some standards in OS, disk media formats, data interchange formats, and peripheral standards, they might get IBM's underbelly the way DEC/UNIX/TAR prevented IBM from cornering the mini market with the Series 1. If Bell had gone with MULTICS instead writing UNIX, the mini market might have gone to something like a Series 1 and EDX environment. Of course, to be competitive in the consumer market, the 68K standard will have to be application object code compatible (Since most micro owners don't get source for every application).