The following message is a courtesy copy of an article that has been posted to bit.listserv.ibm-main,alt.folklore.computers as well.
[EMAIL PROTECTED] (Tom Marchant) writes: > Amdahl was successful in the high end market because it developed > processors using high speed Emitter-Coupled Logic. Dr. Amdahl had > approached upper management at IBM before he left IBM and tried to > convince them to produce such a computer. IBM said that it wouldn't > be profitable. > > Part of the reason they thought it wouldn't be profitable was that in > order to place it on the (liner) price/performance curve, IBM would > have to charge about $10 million for the machine, and at that price > there would not be enough buyers to offset the development cost. > Dr. Amdahl pointed out that they could make a nice profit at a > $4 million price point. IBM was not about to price a high end > processor so inexpensively. re: http://www.garlic.com/~lynn/2007g.html#55 IBM to the PCM market(the sky is falling!!!the sky is falling!!) my observation was competitive offerings/response ... as opposed to what specifically it was that Amdahl was shipping (or why he might have decided leave). The corporation had taken a side-track into the FS project http://www.garlic.com/~lynn/subtopic.html#futuresys which diverted/curtailed a lot of stuff in the 370 product pipeline. part of the reference was to talk Amdahl gave in the early 70s at MIT ... some number of us from the science center http://www.garlic.com/~lynn/subtopic.html#545tech were in the back of the hall ... but didn't say anything ... in any case ... he was asked a number of questions about how he got backing for his company. He mentioned that there was already a few hundred billion dollars in (customer) application 360/370 software ... and even if IBM were to completely walk away from 370 (which might be considered a veiled reference to future system project), there was enuf customer application 360/370 software to keep him in business thru the end of the century. so some of the other part of the business/competitive reference ... i.e. http://www.garlic.com/~lynn/2007g.html#17 The Perfect Computer - 36 bits? http://www.garlic.com/~lynn/2007g.html#23 The Perfect Computer - 36 bits? http://www.garlic.com/~lynn/2007g.html#29 The Perfect Computer - 36 bits? when FS was killed, there was something of a mad rush to get stuff back into the 370 product pipeline. The standard product cycle was 7-8 yrs ... and so the "next generation" stuff (i.e. "811" or 370-XA) wasn't going to be out before the early 80s. So how to revitalize the 370 product line in much fewer yrs ... quickly getting stuff back into development (in parallel with the work going on to produce 370-xa ... in effect two parallel efforts going on at the same time). The answer was the 303x product line. The 370/158 had microengine that ran both 370 emulation code as well as "integrated channel" microcode. The took the 158 microengine with the integrated channel microcode (and w/o the 370 microcode) and called it a "channel director". They then took the 370/158 microengine with the 370 microcode (and w/o the 370 integrated channel microcode) and packaged it as 3031. A standard 3031 might be considered a multiprocessor configuration since it had two 158 microengines ... one dedicated to 370 execuation and one dedicated to integrated channel microcode. The 3032 was a 370/168 repackaged to use one or more 303x channel directors. The 3033 was the 370/168 wiring diagram mapped to newer/faster chip technology. The new chips had something like ten times the circuits per chip ... but was basically only 20percent faster (about the same level technology as being used by Amdahl at the time). However ... just using the same wiring diagram would result in the machine only being about 20percent faster than (3mip) 168-3 ... or about 3.6mips (which was much slower than Amdahl's machine). As a result, there was somewhat of a rush program to redo critical sections of the machine logic to better use more onchip circuits. The result was a machine that shipped around 4.5mips ... rather than 3.6mips. The above reference posts (including some interchange with an engineer that worked in POK at the time) ... also mentioned that a couple of us had co-opt'ed the spare time of some of the processor engineers working on 3033 to help with the design of a 16-way smp machine. everybody thot it was a fantastic machine ... until somebody happened to point out to the head of POK that the favorite son operating system ... wasn't likely to be able to have 16-way smp support for possible another couple decades (definitly not within the lifetime of the specific machine). Then some people were invited to never show up on POK property again. ... in any case, eventually with the 303x out the door, that group then could play "leap-frog" projects with the 3081 group ... and start work on 3090. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
