13 Oct
2006
13 Oct
'06
3:42 p.m.
woodelf <bfranchuk at jetnet.ab.ca> Wrote:
Simon Fryer wrote:
Yes, 24 bit machines did exist. In fact before BSD, there was project Genie
which made an SDS 930 into an SDS 940 (added paging hardware) and made up a
timesharing system. It was the basis of the Tymshare corporation which was
formed in the 60's. The machine had (IIRC) 64k of 24 bit words, paged in
chunks of 4k words. The technique used what they called "relabeling registers"
on the upper 8 bits of the memory address. The software had all sorts of
things, including 'forks'. All before Unix (1969). Quite a system for its
day. It was the competition to the GE 235 (Later 635) machines that started at
Dartmouth. In the early 70's they had a whole bunch of machines in Cupertino
under one roof.
As for strings: They packed 3 to a word (usually).
Trivia: 'BRS 131B' was the system call to crash the system.
Just so you know.
--
Tom Watson
tsw at johana.com
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Heya,
What ever happened to 24 bit words, or is this an easy case?
That is 3 8 bit
char's. :)
I was thinking of Dec's machines and rememberd other brands later.
While 24 bit computers were made, I can't think of any off hand.
> Simon 
16 Oct
16 Oct
11:06 a.m.
On Fri, Oct 13, 2006 at 03:42:30PM -0700, Tom Watson wrote:
Yes, 24 bit machines did exist. In fact before BSD,
there was project Genie
which made an SDS 930 into an SDS 940 (added paging hardware) and made up a
timesharing system. It was the basis of the Tymshare corporation which was
formed in the 60's. The machine had (IIRC) 64k of 24 bit words, paged in
chunks of 4k words. The technique used what they called "relabeling
registers"
on the upper 8 bits of the memory address. The software had all sorts of
things, including 'forks'. All before Unix (1969). Quite a system for its
day. It was the competition to the GE 235 (Later 635) machines that started at
Dartmouth. In the early 70's they had a whole bunch of machines in Cupertino
under one roof.
As for strings: They packed 3 to a word (usually).
Trivia: 'BRS 131B' was the system call to crash the system.
It's funny, I was just thinking about the SDS 940 this week. SIMH has
an emulator for the 940, but so far I have uncovered no software to run
on it.
Has anyone ever found any GENIE tapes or backups?
-Seth
2:33 p.m.
Yes, 24 bit machines did exist. In fact before BSD,
there was project Genie
I have a machine here that's 24 bit -- sort-of. The CPU registers are 24
bits wide, as is the ALU, and the full width of the ALU is used to
calculate memory addresses (24 bit address bus). But the main memory data
bus is 16 bits wide, normally loaded into the bottom 16 bits of the ALU
path, but it's possible to take the bottom 8 bits of memory into the top
8 bits of the ALU path (thus letting you load all the bits of a 24 bit
register from memory in 2 'goes').
I am sure you know what that machine is...
-tony
4:46 p.m.
Tony Duell wrote:
I have a machine here that's 24 bit -- sort-of. The CPU registers are 24
bits wide, as is the ALU, and the full width of the ALU is used to
calculate memory addresses (24 bit address bus). But the main memory data
bus is 16 bits wide, normally loaded into the bottom 16 bits of the ALU
path, but it's possible to take the bottom 8 bits of memory into the top
8 bits of the ALU path (thus letting you load all the bits of a 24 bit
register from memory in 2 'goes').
I am sure you know what that machine is...
Well I don't know off hand. It sounds like a Seymour Cray design for I/O computer
but then you need the 'super computer' to go with it.
-tony
17 Oct
17 Oct
10:39 a.m.
Tony Duell wrote:
No, it's a standalone machine. Hint, most of these machines were 20 bit,
but otherwise identical.
-tony
I have a machine here that's 24 bit -- sort-of. The CPU registers are 24
bits wide, as is the ALU, and the full width of the ALU is used to
calculate memory addresses (24 bit address bus). But the main memory data
bus is 16 bits wide, normally loaded into the bottom 16 bits of the ALU
path, but it's possible to take the bottom 8 bits of memory into the top
8 bits of the ALU path (thus letting you load all the bits of a 24 bit
register from memory in 2 'goes').
I am sure you know what that machine is...
Well I don't know off hand. It sounds like a Seymour Cray design for I/O computer
but then you need the 'super computer' to go with it. 
2:15 p.m.
Tony Duell wrote:
heh heh, wouldn't happen to begin with 'P', would it? :-)
Well I
don't know off hand. It sounds like a Seymour Cray design for I/O computer
but then you need the 'super computer' to go with it.
No, it's a standalone machine. Hint, most of these machines were 20 bit,
but otherwise identical.
6646
days inactive
6650
days old
7 comments
5 participants
participants (5)
-
ard＠p850ug1.demon.co.uk -
bfranchuk＠jetnet.ab.ca -
julesrichardsonuk＠yahoo.co.uk -
sdc695＠yahoo.com -
sethm＠loomcom.com