Re: what's a mainframe, was is Vax addressing sane today

On 9/11/2024 9:21 AM, John Levine wrote:
 > According to Anton Ertl <anton@mips.complang.tuwien.ac.at>:
 >> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
 >>> On Tue, 10 Sep 2024 16:32:05 GMT, Anton Ertl wrote:
 >>>
 >>>> It seems that during the late 1990s, IBM was not particularly interested
 >>>> in mainframe per-CPU performance.
 >>>
 >>> Mainframes were never about CPU performance.
 >>
 >> The S/360 Model 91 and the Model 195 certainly were about the maximum
 >> CPU performance.  And I doubt that IBM would have spent all the effort
 >> with ECL and a superscalar OoO implementation for some of the ES/9000
 >> machines if CPU performance was considered unimportant at the time.
 >>
 >> It's an interesting question why they did not follow up their
 >> superscalar OoO ECL implementations with a superscalar OoO CMOS
 >> implementation in addition to the scalar in-order 9672. ...
 >
 > IBM definitely cared about maximum performance in the 1950s and early 1960s.
Yes.  And remember, one of the goals of S/360 was to provide an architecture that could handle both scientific (i.e. compute bound) and business (i.e. I/O bound) workloads.
 > The goal of STRETCH was specifically to make the fastest possible computer. It sort of
 > succeeded, late and over budget and not as fast as they hoped, but still the fastest
 > computer in the world for a while. It was a success in that they reused a lot of the
 > technology like the fast core memory in later computers.
 >
 > The 360/91 was also intended to be the fastest possible computer, which again it sort of
 > was, late and over budget. One thing that STRETCH and the /91 shared was that they were
 > extremely complicated. STRETCH had variable sized bytes and and addressing modes that I
 > never entirely figured out. The /91 had an instruction queue with loop mode and out of
 > order operations and register renaming and imprecise interrupts. When the CDC 6600 came
 > out, a much simpler design from a tiny company that was nonetheless faster than the /91,
 > they knew they had a problem. The /95 and /195 were minor upgrades of the /91 but that was
 > the end of their supercomputer efforts.
Mostly true, except for the 3090 vector facility.
 > The point of a mainframe is balanced performance. The CPU of a 360/30 was extremely slow
 > but it was fast enough to drive a disk or two and a printer and card read/punch and get a
 > lot of useful work done. Mainframes have had channels since the 709 in the late 1950s so
 > they have a lot of I/O capacity.  Modern ones have terabytes of RAM and exabyte of disk.
Yes.
 > They also care deeply about reliability. Modern mainframes have multiple kinds of error
 > checking and standby CPUs that can take over from a failed CPU, restart a failed
 > instruction, and the program doesn't notice.  I think you'll find a pattern since the
 > CDC shock of making CPUs fast enough to keep the RAM and I/O devices busy while having
 > the error checking and recovery features so the systems keep running for years at a time.
Yes, but they also have to keep producing faster and faster CPUs so they can entice current customers to upgrade and thus meet their revenue goals.
--
  - Stephen Fuld
(e-mail address disguised to prevent spam)