Kahan on the 8087 and designing Intel's floating point

William Kahan, winner of the Association for Computing Machinery's A.M. Turing Award, describes his work as a consultant to Intel to design a new floating point system and incorporate it into the 8087 coprocessor. This design was refined for the 80287 and 80387 chips and integrated into the 80486 (later the 80486DX) processor. This clip is taken from an interview conducted with Kahan by Thomas Haigh for the ACM on 12 March, 2016 in Berkeley, California. Video of the full interview is available as part of Kahan’s ACM profile at amturing.acm.org/award_winner....

[@Alex-jb5tb]

Mr. Kahan also worked on the FPU ? Wow ! I love his Kahan summation algorithm and its many variants so much.

[@nezbrun872]

Awesome. I've never done floating point in hardware, but I have written two floating point libraries in assembly language for the Z80 ~30 years ago, one IEEE754 compliant and the other built for speed.
That was back in the day when as an individual you could understand and he a deep knowledge of how your entire system stack worked, from the keyboard and display down to the wire, and you debugged your code with oscilloscopes and in-circuit emulators.

[@rickburtaine7018]

Yube offers up Kahan so I watch for many reasons. Began as hobbyist with a Z80 system (TDL). Years later I am back with 8080 assembly language. Always wanted to learn the mechanics of floating point. Would you make your Z80 source code available?

[@anthonyrosica5790]

Pure astonishment.

[@logipilot]

Me: barely understanding ieee754

[@nasigoreng553]

Fantastic Video.
Its igniting my reactor to start completing my 8088 / 8087 project.

[@MichaelKingsfordGray]

Is that your real name?

[@AlonzoTG]

I got the bug to build an 8086/8087 but I eventually regained my sanity, good luck tho. =)

[@MichaelKingsfordGray]

HP used BCD floating point. Far superior in every respect to his pure binary.