(j3.2006) [Fwd: [Wg25-full] Fwd: [Numeric-interest] reproducible floating-point results and 754R]

[Van Snyder]

The following is from Bo Einarsson.  It was sent to the IFIP WG 2.5
mailing list.  It may be of interest to you.

Van

-------- Forwarded Message --------
From: Bo Einarsson <boein at mai.liu.se>
To: wg25-full at nsc.liu.se
Cc: wg25-affi at nsc.liu.se
Subject: [Wg25-full] Fwd: [Numeric-interest] reproducible floating-point
results and 754R
Date: Fri, 18 May 2007 08:49:57 +0200
Bo Einarsson
Mathematics Department
Link?pings universitet
SE-581 83 Link?ping, SWEDEN
------ Forwarded message -------


From: David Hough 754R work <754r at ucbtest.org>
To: numeric-interest at ucbtest.org
Date: Thu, 17 May 2007 13:10:50 -0700 (PDT)


For those who have not been following 754R's progress, it's currently
in the sponsor ballot and review phase.      The ballot review
committee
is currently debating the issue of whether reproducible results should
be
available to programmers who want them (and are willing to pay
whatever
performance price is required).

754 said 

 When restricted to a declared subset of the standard,
 these programs should produce identical results on all
 conforming systems

but this was never true, as expression evaluation was never specified.
But most persons whose expertise lies outside the numeric area 
assumed incorrectly that 754 delivered what it promised.    Just a few
of the ways that they were wrong are outlined in Doug Priest's paper

 http://www.validlab.com/goldberg/addendum.html

People who have been active in numerical work sometimes come to
believe
that reproducible results are not possible or even not desirable.

The current draft of 754R is expresses its aims this way:

 For operations specified in the normative part of this standard, 
 numerical results and exceptions are uniquely determined by the
values of the 
 input data, the operation, and the destination, all under user
control.

 Together with language controls it should be possible to write
programs 
 that produce identical results on all conforming systems.

There is some sentiment that this language should be removed, although
it
was part of the 754R working group's charter from the IEEE.

The current draft is not publicly available, but the last public draft
is at

 http://754r.ucbtest.org/drafts/archive/2006-09-17.pdf



I have written some arguments in favor of requiring conforming 
implementations to provide means to obtain reproducible results at

 http://754r.ucbtest.org/msc-ballots/repro.html

I welcome comments from language designers and mathematical software
programmers.     They could also be sent to Bob Davis, bob at scsi.com,
chair of the IEEE MSC which chartered 754R.




My other commentary on 754R ballot reviews is at

 http://754r.ucbtest.org/msc-ballots/ballot3.html

and encompasses other issues beyond reproducibility.

_______________________________________________
Numeric-interest mailing list
Numeric-interest at ucbtest.org
http://mailman.oakapple.net/mailman/listinfo/numeric-interest
_______________________________________________
Wg25-full mailing list
Wg25-full at lists.nsc.liu.se
http://www.nsc.liu.se/mailman/listinfo/wg25-full