[Cfp-interest] WG14 IEEE 754-C binding meeting minutes 2015/02/18

Ian McIntosh ianm at ca.ibm.com
Thu Feb 19 11:01:48 PST 2015


For #3 "Expressions of FP constants done at translation time, yet result is
inexact, so depends upon rounding mode.", the IBM POWER XL and z/OS
compilers have an option -y to control which rounding mode to use for
expressions evaluated at compile time.  Does that meet the requirement?

They also have an option -qfloat=nofold to defer almost all floating-point
expression evaluation until execution time (all but static initialization).
Some (not all yet) operations that cause an overflow or invalid operation
(but not underflow or inexact) are automatically deferred until execution
time (except in static initialization) even without -qfloat=nofold.

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




From:	"Fred J. Tydeman" <tydeman at tybor.com>
To:	"cfp-interest at oakapple.net" <cfp-interest at oakapple.net>, "Rajan
            Bhakta" <rbhakta at us.ibm.com>
Date:	2015-02-18 03:37 PM
Subject:	Re: [Cfp-interest] WG14 IEEE 754-C binding meeting minutes
            2015/02/18
Sent by:	cfp-interest-bounces at oakapple.net



On Wed, 18 Feb 2015 12:01:26 -0600 Rajan Bhakta wrote:
>
>        Fred: Casts are being ignored for floats to integers 'float a =
>(float)DBL_MIN;' for every compiler I've tested. It is very hard to get
>reproducible results as some things are done at translation time and
>others are done at run-time.
>

That should be:

Three problems with reproducible results:
  1. Conversions from FP to integer types done wrong, eg, char c = 1e30;
     Invalid is not being raised.
  2. Casts are sometimes ignored, eg, double d = (float)DBL_MIN; when
     in non-default rounding mode.
  3. Expressions of FP constants done at translation time, yet result is
     inexact, so depends upon rounding mode.


---
Fred J. Tydeman        Tydeman Consulting
tydeman at tybor.com      Testing, numerics, programming
+1 (775) 287-5904      Vice-chair of PL22.11 (ANSI "C")
Sample C99+FPCE tests: http://www.tybor.com
Savers sleep well, investors eat well, spenders work forever.

_______________________________________________
Cfp-interest mailing list
Cfp-interest at oakapple.net
http://mailman.oakapple.net/mailman/listinfo/cfp-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20150219/bbb9f804/attachment.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
Url : http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20150219/bbb9f804/attachment.gif 


More information about the Cfp-interest mailing list