comp.lang.ada
 help / color / mirror / Atom feed
From: robert_dewar@my-dejanews.com
Subject: Re: How smart are Ada compilers?
Date: 1999/03/23
Date: 1999-03-23T00:00:00+00:00	[thread overview]
Message-ID: <7d6quk$njb$1@nnrp1.dejanews.com> (raw)
In-Reply-To: 7d5n3s$mf8$1@nnrp1.dejanews.com

In article <7d5n3s$mf8$1@nnrp1.dejanews.com>,
  Larry Hazel <lhazel@mindspring.com> wrote:
> The Ada 83 code (Apex compiler) I am examining uses
> machine code to access
> 486 block move instructions to copy data into and out of
> memory shared
> between multiple boards.  I would think that an Ada
> compiler targeted to a
> 486 would use such instructions to implement Ada
> assignments of large objects
> without resorting to machine code.  Does Apex (GNAT,
> Aonix, Green Hills, RR,
> etc) generate such code automatically?

Very probably so, (certainly for GNAT the answer is
yes). However, to rely on this would be very bad coding
style. If the board in question requires the generation
of certain machine instructions to work right, then it
is correct (really essential) to use the corresponding
machine insertion, rather than rely on what the compiler
happens to be doing today!

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    




      reply	other threads:[~1999-03-23  0:00 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1999-03-22  0:00 How smart are Ada compilers? Larry Hazel
1999-03-23  0:00 ` robert_dewar [this message]
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox