Re: First language with conditional assignment?

glen herrmannsfeldt <gah@ugcs.caltech.edu>
Mon, 7 Mar 2011 20:11:26 +0000 (UTC)

          From comp.compilers

Related articles
[9 earlier articles]
Re: First language with conditional assignment? compilers@is-not-my.name (2011-03-07)
Re: First language with conditional assignment? compilers@is-not-my.name (2011-03-07)
Re: First language with conditional assignment? compilers@is-not-my.name (2011-03-07)
Re: First language with conditional assignment? neitzel@gaertner.de (2011-03-07)
Re: First language with conditional assignment? bc@freeuk.com (Bartc) (2011-03-07)
Re: First language with conditional assignment? alex.colvin@valley.net (mac) (2011-03-07)
Re: First language with conditional assignment? gah@ugcs.caltech.edu (glen herrmannsfeldt) (2011-03-07)
Re: First language with conditional assignment? robin51@dodo.com.au (robin) (2011-03-08)
Re: First language with conditional assignment? gah@ugcs.caltech.edu (glen herrmannsfeldt) (2011-03-08)
| List of all articles for this month |

From: glen herrmannsfeldt <gah@ugcs.caltech.edu>
Newsgroups: comp.compilers
Date: Mon, 7 Mar 2011 20:11:26 +0000 (UTC)
Organization: A noiseless patient Spider
References: 11-03-006 11-03-023
Keywords: architecture, history, comment
Posted-Date: 08 Mar 2011 15:01:32 EST

(snip previously on conditional expression)
(Then our moderator wrote)


> [Test and set was in the S/360 architecture in the early 1960s. Dunno
> if it was on earlier machines or if they just did it implicitly using
> add to memory. -John]


I believe that many of the earlier core memory machines would
do read-modify-write instructions by delaying the write-back
needed for core memory. (Read is destructive, so an ordinary
read requires a write back.) That would make it impossible
for any other access, even DMA based I/O, to change memory.
(I believe that is how it is done on the PDP-10.)


Also, with S/360 being the first "architecture," designed to
be implemented in technologies, it may have been time to
consider interlocked update separate from the underlying hardware.


-- glen
[Yup, that's what I meant by doing it implicitly. -John]


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.