Nick
Nick
So I finally got around to looking at this . . . and I think you've done the correct thing. Those tests are relativized to the unit roundoff of the...
@mborland : I've approved, but with the caveat that I don't know much about this. We'll probably want a second review from @jzmaddock .
@jzmaddock : How do you build it?
@jzmaddock : Has the module situation improved over the past year? I'm wondering if this PR might work now . .
We'll try again in 2024!
@jzmaddock : Did you forget to merge this one?
> We should allow the user to allocate the memory for this array only once and then run the algorithm as many times as they want, which should fill the...
> Are you convinced now or should I continue talking? I'm convinced.
Yeah, I think getting it to work quickly with multiprecision is the tricky part.
Oh don't forget the `has_denorm`! The trick doesn't work if the floating point unit doesn't support denorms.