Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!mailrus!tut.cis.ohio-state.edu!bloom-beacon!bu-cs!buengc!art
From: art@buengc.BU.EDU (A. R. Thompson)
Newsgroups: comp.arch
Subject: Re: IBM Stretch
Message-ID: <431@buengc.BU.EDU>
Date: 15 Jul 88 20:24:14 GMT
References: <28200172@urbsdc> <11458@steinmetz.ge.com> <4480@medusa.cs.purdue.edu> <2308@sugar.UUCP>
Reply-To: art@buengc.bu.edu (A. R. Thompson)
Followup-To: comp.arch
Organization: Boston Univ. Col. of Eng.
Lines: 24

In article <2308@sugar.UUCP> karl@sugar.UUCP (Karl Lehenbauer) writes:
Deleted stuff
>Yeah, the 360/90 was the machine that IBM was more or less forced to build.
>As I understand it, IBM didn't think there was a scientific supercomputer
>market and was quite surprised when CDC received a bunch of orders for

When Gene Amdahl first proposed his high speed box idea he was with IBM.
They rejected the idea on the grounds you state.  He then said he would do
it on his own.  The brass tried to stop on the grounds they thought he
would fail, and they liked him too much to see him go down the tubes.
Amdahl tells the story of one of the bigshots putting his arm around him
and begging him not to ruin his career.  Well, he took the chance and the
rest is history.

This is a recollection from an article about Amdahl that appeared in one
of the trades (I think it was Datamation but can't remember) about eight
years ago.

>the then-new 6600.  Consequently, IBM announced a vaporware product, a high 
>speed scientific supercomputer, the 360/90, causing most 6600 orders to be 
>cancelled, prompting CDC's antitrust suit against IBM.  IBM settled out of
>court with CDC and did finally build the 90, but it was years late and did
>not achieve its original stated performance.
>-- 
>-- karl@sugar.uu.net aka uunet!sugar!karl