19 Oct
2011
19 Oct
'11
5:47 p.m.
[Should have changed Subject: header that sooner, but it will screw
threading at this point. --rma]
From: Dave McGuire
Sent: Tuesday, October 18, 2011 1:44 PM
On 10/18/2011 04:15 PM, Rich Alderson wrote:
> When IBM wanted to port the IBM OS family to
POWER, they did the same thing.
> (Z, in your list above.)
Hey Rich...If you're suggesting z/Series
machines are based on POWER
processors, please provide some references. Everything I've ever seen
about them (including rooting around inside the z890 in the next room)
says they're completely self-contained with their own CPU implementation
family. Lots of people have suggested that modern IBM mainframes are
somehow built around RS/6000 chipsets with different microcode, but have
never been able to back up those claims.
(not trying to be argumentative, and would happily
stand corrected,
but I keep hearing people make this statement and then fall silent when
asked to prove it)
Hey, Dave,
It's not a deeply researched statement. It's based on reportage in the
trade press back when the z-, i-, and was it pSeries(?) systems started
coming out. It sounds like emulation to me.
Also, please note that I used "POWER" as a generic, not distinguishing
between the chip set in a PC/RT and those in Watson's blades. I *do*
distinguish "POWER" from "POWERpc", since there are specific
differences
that make the latter the choice for a small system. So if it's POWER6
or POWER7, or whatever, I'm not making that much of a distinction.
I don't think anyone has seriously claimed that "RS/6000 chhipsets with
different microcode" are involved. There's been a lot of water under the
dam since those came out. My understanding is that a multi-level system
is in use, with nanocode to create the microcode engine to run the code
that makes it look as though you're on hardware compatible with a 360 from
1968, with bums and hacks to bring it past the 31-bit extension of the
24-bit address space into the 64-bit present day.
If that's not correct, fine. I'll withdraw my comment. But that's the
impression given by IBM's own press releases going back to the introduction
of zSeries, and that's whence I drew my conclusions.
Thanks,
Rich
Rich Alderson
Vintage Computing Sr. Server Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/