[email protected] (John McKown) writes:
> Look at the current "System i". It really looks a lot like what has been
> discussed here. It has, theoretically, a 128 bit virtual addresses.
> Everything is an object. It has the single level storage so there aren't
> really any "disk files" as such. And once an address has been associated
> with an object, that address will _never_ again be used for some other
> object. So when an object is freed, its address becomes unavailable. Thus
> it is impossible to access a freed object or accidentally use an old
> address to access a different object.

re:
http://www.garlic.com/~lynn/2014c.html#71 assembler
http://www.garlic.com/~lynn/2014c.html#74 assembler

folklore is that when future system failed, several people retreated to
rochester and did (a "simplified" as) the s/38, capability, single level
store, etc. 
http://www.garlic.com/~lynn/submain.html#futuresys

One of the simplification was treating all disk as common
allocation pool with scatter allocation. As a result, the complete
infrastructure had to backed up as single operation ... and any single
disk failure resulted in having to make a complete restore of everything
... restore after a single disk failure could take a day or more.  As a
result, s/38 was early adopter of RAID ... patent issued to San Jose
disk engineer late 70s.
http://en.wikipedia.org/wiki/RAID

AS/400 was follow-on to combination of s/38 & s/36 ...  single-level
store retained ... 
http://en.wikipedia.org/wiki/IBM_System_i
but capability-based address was dropped.
http://en.wikipedia.org/wiki/IBM_System_i#History

circa 1980 there was an effort to replace large number of internal
microprocessors with 801/risc Iliad chips (4361/4381 followon to
4331/4341, as/400, lots of controller microprocessors, etc) ... for
various reasons ... those efforts floundered and they did continued
business as usual with customer cisc chips.

later in the 90s, the as/400 was part of the AIM (apple, ibm, motorola)
group doing power/pc (single chip 801/risc) activity and finally moved
to 801/risc (now system i).

past posts 801, risc, illiad, romp, rios, power, power/pc, etc
http://www.garlic.com/~lynn/subtopic.html#801

as an aside, after the failure of the 1980 effort to move to 801/risc
... some number of the chip engineers left and show up on risc projects
at other vendors.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to