Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!uunet!husc6!rutgers!ucla-cs!zen!ucbvax!hplabs!hp-pcd!uoregon!omepd!randys From: randys@mipon3.intel.com (Randy Steck) Newsgroups: comp.sys.ibm.pc Subject: Re: Intel Microprocessors Message-ID: <940@omepd> Date: Fri, 7-Aug-87 20:44:27 EDT Article-I.D.: omepd.940 Posted: Fri Aug 7 20:44:27 1987 Date-Received: Sun, 9-Aug-87 11:31:28 EDT References: <1112@lznv.ATT.COM> <399@aucs.UUCP> <3225@cucca.columbia.edu> <789@unccvax.UUCP> <1924@Shasta.STANFORD.EDU> Sender: news@omepd Reply-To: randys@mipon3.UUCP (Randy Steck) Organization: Intel Corp., Hillsboro Lines: 50 After due deliberation, I just have to say something. I think that this discussion is really missing several vital points. First, the reason that IBM chose the 8088 in the first place for the PC was because they could build a PC around it without even trying hard. It just was not possible at the time to get interface chips for the 68000 or any other 16-bit processor. This, along with the 8-bit data paths, allowed IBM to make PCs very cheaply and to essentially take over the PC market. In other words, they made a good marketing decision. Second, IBM has used 68K processors in equipment that they have built for external sale as well as internal use. An example of this was given in a Byte article a couple of years ago. It was a measurement system using the 68K. I think that the conclusion that any perceived similarity between the 370 and the 86 family justifies the use of the 8086 is unreasonable. Third, segments are a reasonable programming paradigm. Intel has said this before to the derision of the world at large. The most that one can fault Intel for is the choice of size for these segments (but what could one reasonably expect in 1978?). Larger segments, as found on some mainframe computers, would simplify the task of programming the part. Memory management (a la 286) is a fall-out from the segmented addressing of the 86 architecture. Fourth, I must disagree with the statement that the introduction of the PS/2 was an opportunity missed by IBM to "get out of the tar pit of Intel processors". Would you really like to see a system introduced as a follow-on to the PC that would not run all of the code already written for the PC? I am sure that the marketeers at IBM get a big chuckle every time someone suggests such an idea. I am a design engineer at Intel working on 32-bit processors and math coprocessors. I personally believe that there are things that could be better in the 86 architecture, but please don't assert that Motorola did everything right in their processors either! If architectures are as perfect as the marketing people say they are (those for any company, not just Intel and Moto), then why do newer parts keep getting designed? Performance is only one of the reasons. If you really want an Intel processor to flame, then you should learn about the 432. (Yes, I worked on that one too). It was not a commercial success but served its purpose. I can't wait until I get a 386/486 on my desk. Then I won't have to rely on the silly little VAX for everything. Randy Steck Intel Corp. Disclaimer: Do you really think Intel agrees with anything I say?