Fw: Useless sNaNs... or useful?

[Ian McIntosh <ianm@xxxxxxxxxx>]

"I used to use a pre-754 system ..." does not refer to any companies doing that today. It refers to the Control Data 6600, 7600 and their successors that I used to use. The memory initialization pattern I referred to was called DEBUGA. There was also DEBUG, without the addresses, but DEBUGA was more useful because you could find out which variable or array element it was from. (BTW, two of 754's features (Infinites and NaNs (6600 Indefinites)) and part of their propagation riles appear to have evolved from the 6600 and 7600, maybe via Cray.)

By "I know of companies doing this themselves today", I meant those using computers, not computer vendors. That they're doing it is likely not a secret, since one shares software with others, but I don't have their permission to name them and I don't think their names are relevant.

The point is that Signaling NaNs are useful to some and being used even if their representation is not totally portable. Maybe the 754 committee accomplished more on this than thought. 8<)

Another way of looking at it: Would Interval Arithmetic be useless if its representation was not totally portable?

As far as C99 support, C99 left out a lot of 754-1985, and now there's 754-2008. The C Floating Point "discussion group" is working to prepare a Technical Report to the C Standard Committee to improve C's support for 754. Adding support for Signaling NaNs is on the list of upcoming topics. Maybe C 201x will have it, or maybe not.

- Ian McIntosh IBM Canada Lab Compiler Back End Support and Development

----- Forwarded by Ian McIntosh/Toronto/IBM on 10/15/2010 03:27 PM -----

From:	Dan Zuras Intervals <intervals08@xxxxxxxxxxxxxx>
To:	Ian McIntosh/Toronto/IBM@IBMCA
Cc:	stds-1788@xxxxxxxxxxxxxxxxx, Dan Zuras Intervals <intervals08@xxxxxxxxxxxxxx>
Date:	10/15/2010 03:20 PM
Subject:	Re: Fw: Useless sNaNs... or useful?
Sent by:	ben@xxxxxxxxx

---

> Subject: Fw: Useless sNaNs...  or useful?
> To: stds-1788@xxxxxxxxxxxxxxxxx
> From: Ian McIntosh <ianm@xxxxxxxxxx>
> Date: Fri, 15 Oct 2010 14:25:31 -0400
> 
> 
> 
> . . .
> 
> One important use is to initialize all floating point variables to
> signaling NaNs.  If they are inadvertently not properly initialized lat=
> er,
> you will be told if the program attempts to use them.  I used to use a
> pre-754 system that had an option to initialize everything to their
> equivalent of a negative signaling NaN with the payload containing that=
> 
> particular memory address and a few other bits to make it likely to fai=
> l if
> used as a pointer or subscript.  That found bugs even in some carefully=
> 
> written programs that "worked" on other systems (just not correctly).  =
> I
> know of companies doing this themselves today.  It would be a worthwhil=
> e
> feature in every system, and doesn't depend much on NaN portability.  I=
> t is
> affected by the second imperfect portability - whether the payload is o=
> r is
> not preserved.
> 
> . . .
> 
> - Ian McIntosh   IBM Canada Lab         Compiler Back End Suppor=
> 

		 I know of no companies doing this today.

		 Please name them.


		 		 		 Dan