Re: PDP-11 OS Release Dates
From: Brian Inglis (Brian.Inglis_at_SystematicSw.ab.ca)
Date: 07/29/03
- Next message: Colin Blake: "Re: Mozilla... the new NULL process"
- Previous message: Bill Todd: "Re: Forbes cover story: HP Merger Successful"
- In reply to: John Sauter: "Re: PDP-11 OS Release Dates"
- Next in thread: Bill Gunshannon: "Re: PDP-11 OS Release Dates"
- Reply: Bill Gunshannon: "Re: PDP-11 OS Release Dates"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 29 Jul 2003 09:57:22 GMT
On Tue, 29 Jul 2003 02:44:25 GMT in alt.sys.pdp11, John Sauter
<J_Sauter@Empire.Net> wrote:
>cstacy@dtpq.com (Christopher C. Stacy) wrote:
>
>RAM...that's disk, right? (As opposed to the sequential access memory
>provided by magnetic tape.)
>
>John Sauter responded:
>
>No, RAM means memory whose access time does not depend on
>the address of the previously accessed location. Neither disk nor
>tape qualify, but core and semiconductor memory do.
Neither do delay lines or drums, but Williams tubes do.
>Before the invention of core, computer memory was often implemented
>using a disk. When an instruction completed its execution, the CPU
>would wait until the address of the next instruction rotated around to
>the read head so it could fetch that instruction. This wasted time
>could be avoided by placing instructions in memory such that the
>next instruction was about to reach the read head when the previous
>instruction was completed. It was before my time, but I have been
>told that the IBM 650 had an assembler called SOAP that would
>do this optimization for you. I count SOAP as the first high-level
>computer language.
The IBM 650 was the Magnetic Drum Calculator/Data Processing
Machine, designed for punched card processing. Options included
the 653 Auxiliary High Speed Storage Unit (core), 355 Disk
Storage Units, and 727 Magnetic Tape Units.
Single words of instructions or data were fetched from the drum
and executed or operated on, or written to the drum. This is what
required the timing optimizations.
Tapes and disks were used to load blocks of instructions or data
into core, where they were executed.
Drum timing optimization was done by hand on machine code using
(female) clerical staff originally, then as symbolic assembler
codes became useful, timing optimization was usually included in
assemblers for drum machines.
These assemblers were no more an HLL than any optimizing
assembler created more recently to use shorter instruction forms
where possible or schedule instructions to avoid resource
conflicts and improve performance.
The first generally recognized HLLs appear to be the Autocodes
available on various machines.
>The invention of core memory made SOAP (and the IBM 650)
>obsolete.
Core was available on the 650 from 1955 (all of 60 words == 1
disk track), and they were built at a rate of one a day from
1956, until the last of nearly 2000 650s was manufactured in
1962.
http://www-1.ibm.com/ibm/history/exhibits/650/650_intro.html
Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
--
Brian.Inglis@CSi.com (Brian dot Inglis at SystematicSw dot ab dot ca)
fake address use address above to reply
- Next message: Colin Blake: "Re: Mozilla... the new NULL process"
- Previous message: Bill Todd: "Re: Forbes cover story: HP Merger Successful"
- In reply to: John Sauter: "Re: PDP-11 OS Release Dates"
- Next in thread: Bill Gunshannon: "Re: PDP-11 OS Release Dates"
- Reply: Bill Gunshannon: "Re: PDP-11 OS Release Dates"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
