CG-10-08.md
October 8, 2024 · View on GitHub
Agenda for the October 8th video call of WebAssembly's Community Group
- Where: Virtual meeting
- When: 2024-10-08, 16:00-17:00 UTC (2024-10-08, 9am-10am PDT, 18:00-19:00 CEST)
- Location: link on W3C calendar or Google Calendar invitation
Registration
No registration is required for VC meetings. The meeting is open to CG members only.
Agenda items
- Opening
- Proposals and discussions
- Update on and (possible) phase 2 vote for 128-bit-arithmetic (Alex Crichton, 30 minutes)
- Closure
Agenda items for future meetings
None
Meeting Notes
Attendees
- Hao Tran
- Derek Schuff
- Paul Dennis
- Zalim Bashorov
- Jamey Sharp
- David Thompson
- Aruokhai Joshua
- Jeff Charles
- Hunter Demeyer
- Saul Cabrera
- Alex Crichton
- Luke Wagner
- Yury Delendik
- Paolo Severini
- Nick Fitzgerald
- Manos Koukoutos
- Andrew Brown
- Nuno Pereira
- Chris Fallin
- Francis McCabe
- Julien Pages
- Adam Bratschi-Kaye
- Brendan Dahl
- Jake Enget
- Emanuel Ziegler
- Jakob Kummerow
- Oscar Spencer
- Heejin Ahn
- Mattias Liedtke
- Sam Clegg
- Ryan Hunt
- Bailey Hayes
- Ilya Rezvov
- Daniel Lehmann
- Chris Woods
Proposals and discussions
- Update on and (possible) phase 2 vote for 128-bit-arithmetic (Alex Crichton)
AC: presenting slides
Discussion:
BH (chat): Thank you for the overview!
ZB(chat) : nice overview
RH(chat): Agree, thanks for the thorough walk through of the possibilities here.
LW(chat): Great analysis of the alternatives
HD: high level question: is there a reason you didn't go for FP128?
AC: it was unrelated to the original motivation
HD: in the scientific world, higher precision FP are important and could help for portability of scientific code.
AC: I’m not very familiar with that world, i’m not sure I would be a good person to push that forward.
HD: I think this is still good for what it’s for.
Poll to advance to phase 2 and rename to “wide-arithmetic” SF: 14 F: 16 N: 1 A: 0 SA: 0
JK (chat): big +1 to mul_wide_{s,u} instead of mul128. I'm still not sold on add128 being superior to add_with_carry, but (1) I'm also not sure it's worse, and (2) we can continue that conversation elsewhere :-)