[Cfp-interest 2919] JTC1/SC22/WG14 meeting day 1 summary

[Rajan Bhakta]

Hello,

Here is a summary of the first day of the WG14 meeting. I also am forwarding some more editorial comments brought forward by another WG14 member

Action items (for CFP):
  For TS-18661 part 5, for FP_{FLT,DEC}_EVAL_METHOD "Use of unsupported values of width" UB, make it diagnosed rather than UB. Ex. As a constraint violation (needs a new constraints section).
  For TS-18661 part 4, the library features should be split into different headers, remove WANT macros (users can use __has_include and #include), and to review the names (Ex. add reduc_ and scaled_ to the reserved prefixes list).
  Look into timing of these changes to both parts see if they can be made before the ISO deadline.

Decisions:
  WG14 agrees to bring TS 18661-5 v2 to ballot including editorial revisions documented as of 2023/10/16.

Upcoming:
  TS 18661-4 v2 ballot question will be answered tomorrow depending on the results of the timing of the ISO deadlines.
  We will probably need to have a 2 week WG14 review period for any significant changes to any of the parts before it can go forward for ballot.

Regards,

Rajan Bhakta

From: Shivashankar Maddanimath <shivashankar.maddanimath at gmail.com>
Date: Monday, October 16, 2023 at 10:10 AM
To: Rajan Bhakta <rbhakta at us.ibm.com>
Subject: [EXTERNAL] Comments on n3163 - ISO/IEC TS SPECIFICATION 18661-4
Hi Rajan, Pls find my comments below, I am not so sure about the comments on Nan, pls ignore if these are invalid. Page vii: Paragraphs 25 and 30 are duplicates Page 4: Paragraph 25 Description of reduc_sumabs Sum of infinite with different
ZjQcmQRYFpfptBannerStart
This Message Is From an Untrusted Sender
You have not previously corresponded with this sender.
    Report Suspicious  <https://us-phishalarm-ewt.proofpoint.com/EWT/v1/PjiDSg!12-vrJCw6vGy2k9Yc8-OYke6d-BLnPQ5xWXPyIaU7JZCdwZLhLufsM9E9f_2OIawOyMkwi_cnlOU1DXvKGXn2TtabWNmWYHikeznXpDqg2uHIodq33gIGh4fCQ$>   ‌
ZjQcmQRYFpfptBannerEnd
Hi Rajan,

Pls find my comments below, I am not so sure about the comments on Nan, pls ignore if these are invalid.

Page vii: Paragraphs 25 and 30 are duplicates
Page 4: Paragraph 25 Description of reduc_sumabs
             Sum of infinite with different signs is missed.
             In statement "if any member of array p is a quiet NaN, the functions return a quiet NaN." shouldn't this be just Nan ?
Page 5: Paragraph 25 Description of reduc_sumsq
             In statement "if any member of array p is a quiet NaN, the functions return a quiet NaN."  shouldn't this be just Nan ?
Page 6: Paragraph 30 Description of reduc_sumprod
             In statement " The reduc_sumprod functions compute the dot product of the sequences of members of the arrays p and q: ∑ (��[��] × ��[��]) ��−1 ��=0 ."
           This is summation of dot products of corresponding array members from p and q but wording states it computes dot product which means its a dot product of all p and q members.

Regards,
Shiv Maddanimath



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20231016/8d6489ba/attachment-0001.htm>