The Great Debate @ARITH23: John Gustafson and William Kahan

Channel:dancefan · Length:1:34:42 · Published:2016-07-30 · Views:4223

Watch on YouTube

Summary

A 95-minute recorded debate at the ARITH23 computer arithmetic conference between John Gustafson, proponent of unums/posits, and William Kahan, architect of IEEE 754 floating point. They argue the merits of the new number formats versus conventional floats and interval arithmetic.

Why HN recommends it

It is the go-to primary source whenever HN discusses posits versus IEEE floats, valued for Kahan's sharp critique despite the poor audio.

Audio quality is not great but there are links to slides from Gustafson and Kahan — HN, 2017

Math & theory (other) · Computer architecture & CPUs

Subtitles (excerpt)

and as you know John guson recently published a book called The End error in which he proposes a new floating Point form that called UNS which is short for Universal number and as a as AOS replacement for conventional FL and as the book's title suggests he makes some rather strong claims about what UNS can do so for example that they're easier to use than floating Point more accurate less demanding on energy on memory and bandwidth or power and guaranteed to give bitwise identical results and in most situations faster so in many places throughout the book John acknowledges the contributions of William Khan floating point R of course and so and uses many of Khan's examples throughout the book uh about how floating Point can go wrong to illustrate the potential advantages of youo so for example one section of gusterson's book is entitled The Wrath of K another section entitled another con so indeed the great debate is foreshadowed in section 18.1 where gusterson responds at length to a challenge proposed by Professor Khan to an early drafted the book now uh this development has not stopped with this and John has continued to develop these ideas and most recently he's proposed something called sorin set of

On Hacker News

The Great Debate @ARITH23: John Gustafson and William Kahan [VIDEO] Audio quality is not great but there are links to slides from Gustafson and Kahan — celias, 2017-03
A related discussion about number types: The Great Debate @ARITH23: John Gustafson and William Kahan [VIDEO] Kahan points out disadvantages of interval maths. — porkbrain, 2023-04
Original posits are variable width, making them nearly useless for high performance parallel computations. Later versions don't add anything of use for low precision neural networks, and lack of hardware support anywhere make them too slow for anything other than toying around. See also [LINK] and [VIDEO] — SideQuark, 2023-01
I did some reading on unum a while back, inc. Kahan's criticisms. There's another side. There's a vid on youtube with Kahan and Gustafson talking over it. Kahan goes in like a wolf onto a lamb, much to gustafson's shock and hurt. Gustafson points out some of Kahan's criticisms are plain wrong, even citing the page in the book, then says something like "we should be working together on this, why a… — zasdffaa, 2022-09

Related videos