[Cfp-interest] WG14 IEEE 754-C binding meeting minutes 2018/06/21
Rajan Bhakta
rbhakta at us.ibm.com
Thu Jun 21 10:01:18 PDT 2018
Attendees: Rajan, Jim, Ian, Fred, David H.
New agenda items:
None
Carry over action items:
Ian: See if there is an incompatibility between C and C++ for
constants being evaluated to a wider format (Ex. FLT_EVAL_METHOD affects
constants in C++, and wider return values) - Keep open (Hubert: Not
defined and left up to C)
Jim: Update the binding table in parts 1 and 2 to handle the new
IEEE-754:2018 functions when published. - Keep open.
Fred: See where we are for inconsistent spec for infinities - Done.
Last meeting action items:
David: Check the min/max C specification to ensure it matches what
IEEE has. - Not done.
David: Check the augmented* C function specifications to ensure they
match what IEEE has. - Not done.
Rajan: Give a valid version of C17 to Jim for posting on CFP. - Done.
Jim: Ask Blaine about missing references to DECIMAL_DIG (only one
found was the example). - Done.
Jim: Make the quantize and general canonical results changes as
editorial changes for part 2 as per Jim's 2018/05/12 note. - Done.
Jim: Create a CR for llquantexp with the fix in Jim's 2018/05/12 note.
- Done.
Jim: Review specification for raising the inexact exception in the
TS's. - Done.
Jim: Put CR21 on the agenda for next time. - Done.
Jim: Get N documents for the features mentioned in
http://wiki.edg.com/pub/CFP/WebHome/C_support_for_754-201x-20180518.pdf
and use this as new document for Pittsburgh. - Done.
Rajan: Reply to Joseph's note and ask for any particular concerns. -
Done.
New action items:
Jim: Remove screen-share information from the agenda.
Jim: Missing line numbers for part 2 (cfp2x-20180612.pdf)
Jim: Specification for inexact: Make the editorial change as given in
the modified agenda with links.
Fred: Fix the wording for CR21 to specify odd significand.
Jim: C support for 754-201x: Say as per current IEEE draft and add in
the N document numbers for 9.5 and 9.6 additions.
Jim: Submit the augmented arithmetic and min/max spec for 754:201x
binding proposals to WG14 after removing change marks.
Jim: Create a CR for changing "not infinite" to "finite" as per the
description for remainder (18661-1, F.10.7.2).
Jim: Create a paper/proposal for the loosening of the totalorder*
specifications to match 754-201x.
Fred: Recheck 'Functions and infinity' issues with 754 draft 238.
Jim: Create a CR to change CFP to say for Annex F of C that
pow(+/-0,-inf) has no exception (i.e. Remove the latitude to raise
div-by-zero).
Jim: Create a CR to change the reduction operations description from
should->shall
Next Meeting(s):
Wednesday, July 25th 2018, 11:00 EDT, 8:00 PDT
Same teleconference number.
Discussion:
IEEE 754 revision:
David: Finding more things that are incorrect or inconsistent.
Various contributions from the C FP group (this group).
C++ liaison:
Nothing.
Ian: Looking into getting a connection with the C++ committee.
Action item details (
http://wiki.edg.com/pub/CFP/WebHome/CFP_meeting_agenda_20180621-links.pdf
):
quantize and canonical results changes for part 2 (and part 1):
Fred: Since these are changes to the main body of the standard
other than an annex it is not good.
Jim: Since Decimal can only be 754 it doesn't need to be in Annex
F.
Fred: It should mention domain error as well as invalid.
Jim: It does.
llquantexp:
Since it is not a 754 function, we need a CR. It has been
submitted. Blaine should assign a CR number for this.
Missing line numbers in part 2.
Specification for inexact:
Fred: Should overflow and underflow be in quotes as well?
Jim: Considering them as conditions.
David: underflow could be exact so don't need to raise the
inexact exception.
Consider this as an editorial change.
CR21:
Jim: It is for the significand being odd. We should say that.
David: Perhaps say when both the significands are odd?
Jim: Yes, a little strange though since significands are usually
expressed as fractions.
Jim: both choices are odd -> both choices have an odd least
significand digit, and similar for the end of the last line.
N docs for 754-201x support:
Jim: Got the N numbers.
Added tanpi to the list.
Need to say as per current IEEE draft.
Need to add in the N document numbers for 9.5 and 9.6 additions.
Augmented arithmetic:
Jim: Added in the invalid exception for augadd, augsub
min/max spec:
Jim: Added in part 3 changes.
This is a proposal for changes for TS's 1-3, not a CR. It is at
the same time as support for the next version of the 754 standard.
Consistency for infinities (note chain titled Functions and
infinity):
Changing not infinite to finite as per the description for
remainder.
totalorder*: Now up to the implementation on how to order them
done on purpose by 754. Should make this a paper for a new feature/change,
not a CR.
Although we wish we had originally had NaN refer to both quiet and
signalling, it is too much of a change now (with the potential of missing
areas or having errors) so we will not do it.
rootn(1, 0): To match pow, rootn(1, 0) should give 1.
David: In 754:2008, n = 0 is invalid for rootn so changing that
would be a substantive change.
If something changes, it would have to be in 754.
rootn(inf, n) missing in 754:
David: Added in the latest 754 draft 238.
pow(0,-inf):
Change CFP to match 754. New CR.
powr:
David: The whole idea of these functions was to simplify
exception processing.
Problematic for which sub-exceptions should be signalled.
David: CFP shouldn't change since the same problem for both 754
and CFP.
reduction operations need should -> shall.
Other issues:
None.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20180621/cbf41fcc/attachment-0001.html
More information about the Cfp-interest
mailing list