comp.lang.ada
 help / color / mirror / Atom feed
From: Robert Dewar <dewar@gnat.com>
Subject: Re: REaltime ADA
Date: 1999/11/07
Date: 1999-11-07T00:00:00+00:00	[thread overview]
Message-ID: <802ir6$9cv$1@nnrp1.deja.com> (raw)
In-Reply-To: 3824BBA3.E5E636CC@acm.org

In article <3824BBA3.E5E636CC@acm.org>,
  Bo Sanden <bsanden@acm.org> wrote:
> Another question is about Ada real-time performance. The
> features are usually intended to run efficiently on a bare
> machine. Running on top of something else, they may not
> provide acceptable real-time performance, depending of course
> on what your performance requirements are.

That's too narrow, the features in Ada 95 are definitely
intended to be compatible with a wide range of operating
environments. In practice these days, at least in our
experience, most real time Ada programs are NOT running
on bare machines, but instead are using real time kernels
such as VxWorks, Lynx, RTEMS (we have customers successfully
running high performance real time programs in such
environments). I think this is pretty much the most standard
approach these days. What has happened is that machines have
got fast enough so that the extra overhead that comes from
using one of these kernels is acceptable, and of course there
are lots of other advantages in abstracting oneself one step
from the bare board. For example, in the case of GNAT, the
tasking interface from the Ada runtime to VxWorks is identical
for all versions of VxWorks.

It's certainly possible to run on bare boards, and we do have
customers doing so, including one very interesting application
involving internet switches. However, that seems to be the
exception these days.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Before you buy.




  reply	other threads:[~1999-11-07  0:00 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1999-11-04  0:00 REaltime ADA Justin CHampion
1999-11-04  0:00 ` Matthew Heaney
1999-11-04  0:00 ` Ted Dennison
1999-11-04  0:00 ` Marin Condic
1999-11-06  0:00 ` Bo Sanden
1999-11-06  0:00 ` Bo Sanden
1999-11-07  0:00   ` Robert Dewar [this message]
1999-11-07  0:00 ` tmoran
1999-11-12  0:00 ` A.Subbotin
1999-11-12  0:00 ` A.Subbotin
replies disabled

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