Kristoffer Carlsson
Kristoffer Carlsson
The most helpful thing would be to read the docs I have written, comment on that and try the macro out for yourself to see how you think the workflow...
Regardless of `NaN`s, it would be good to figure out why `Matrix` and `SMatrix` differ. I guess just peppering some print statements throughout the code should reveal where things diverge.
With StaticArrays 0.12.4 I get the same result between static matrix and matrix running the example in the first post. With StaticArrays 1.2.13 I get your result. So something happened...
> Only the real part of a dual number is Real. Subtyping relations in Julia should not be confused with the relations of the number sets themselves. Subtyping in Julia...
As long as the required methods are implemented for the custom type there should be no problem using ForwardDiff with it.
Why did you remove the stacktrace?
On this PR there are a bunch of `nsz contract` fastmath flags on the operations? Any idea where those come from?
> Try using Chunk(7) instead of Chunk(8), like in the example. How do you measure this? Just so we do the same.
I meant how you measure the number of instructions. Are you using Cthulhu to step in or just directly calling the function with dual numbers?