News for hackers. Stuff that matters.

A comment on An In-Depth Look at Google's Tensor Processing Unit Architecture. Parent comment

Re:An In-Depth Look at Google's Tensor Processing Unit Archi... by gcr on Wednesday April 05, 2017 @03:27PM (#14044805)
Yeah...you could implement 16-bit multiplication/addition as two 8-bit multiplications plus carry... so worst case, if you want 16-bit multiplies, you implement it yourself
[ Parent | on HN ]
Re:An In-Depth Look at Google's Tensor Processing Unit Archi... by stephencanon on Wednesday April 05, 2017 @04:54PM (#14045528)
Four multiplications, not two (three if you only need the low 16b of the result).
[ Parent | on HN ]
Re:An In-Depth Look at Google's Tensor Processing Unit Archi... by yorwba on Thursday April 06, 2017 @03:57AM (#14048758)
You can get the full 32-bit result in only three multiplications using https://en.wikipedia.org/wiki/Karatsuba_algorithm , but it requires more additions.
[ Parent | on HN ]
Re:An In-Depth Look at Google's Tensor Processing Unit Archi... by stephencanon on Thursday April 06, 2017 @07:56AM (#14049738)
Yeah, in practice Karatsuba isn't a performant option for small operands unless your multiplier is catastrophically slow. (And it still doesn't get you to two multiplications.)
[ Parent | on HN ]