Matrix Division on the 28C/S (and later)

[Werner]

(03-04-2020 04:39 PM)Albert Chan Wrote:  DotFMA probably will not gain much accuracy.
Example, if we fuse for every step, with tiny a, such that a+1=1

DotFMA([a 1 -1], [1 1 1]) = (a+1) - 1 = 1-1 = 0

However, fusing result at the end help (essentially "double" precision until final sum)
see CompDot, CompDotFMA, http://rnc7.loria.fr/louvet_poster.pdf

Well, yes, that's the idea: don't use the FMA for the dot product itself, but for the correction term, a bit like Kahan summation, but for dot products.

Cheers, Werner