Mathematicians discover a perfect way to multiply (2019)

Re:Mathematicians discover a perfect way to multiply (2019) by thelazydogsback on Wednesday July 22, 2020 @03:18PM (#23920555)

This is amazing to me: > "Their conjecture was based on a hunch that an operation as fundamental as multiplication must have a limit more elegant than n × log n × log(log n)"

Only in maths (and sometimes in its "implementation"/ expression in nature in conjunction with physical systems) can such a subjective viewpoint be so powerful.

Also blows me away, at least at first, that multiplying using an FFT would be the fastest method so far. (I guess I have an older mindset based on how slow FFTs used to be.) I wonder in actual implementations at how many digits the perf curves meet at break-even compared to other methods.

[ Parent | on HN ]

Re:Mathematicians discover a perfect way to multiply (2019) by gowld on Wednesday July 22, 2020 @03:40PM (#23920754)

When you say FFT is slow, did you know any way to multiply large numbers that was fast? Or you assumed it was without evidence?

[ Parent | on HN ]

Re:Mathematicians discover a perfect way to multiply (2019) by thelazydogsback on Wednesday July 22, 2020 @04:32PM (#23921249)

Huh?

[ Parent | on HN ]

Re:Mathematicians discover a perfect way to multiply (2019) by perl4ever on Thursday July 23, 2020 @12:11AM (#23924107)

I completely missed this when I first looked at the link, but based on another comment, the break even point is considerably beyond the size of any number that would fit in the universe as we know it. Or any number you could write down if you had a universe for each atom in our universe, etc.

...so it's not really practical.

[ Parent | on HN ]