On Tuesday, 1 November 2016 at 22:22:17 UTC, Ivan Kazmenko wrote:
On Tuesday, 1 November 2016 at 22:13:29 UTC, e-y-e wrote:
...
Ouch, and I completely misread the TL;DR paragraph, sorry for
the noise!
My fault, note to self: always re-read everything.
It's logical to have cumulativeSum specialized from
cumulativeFold when sum is a specialization of fold.
I have yet to write a program which relies on Kahan summation,
but I can imagine that being useful.
To be entirely honest, I haven't experienced any _problems_ with
the accuracy of naive summation. But Kahan summation does improve
accuracy in some cases and I want that assurance.
It does have its own inaccuracies though and I think (may be
wrong, again, its late) a flag like -ffastmath would probably
eliminate any accuracy gain.