Closed mariuszhermansdorfer closed 1 year ago
Regarding the primid, lets discuss on this thread https://github.com/cadop/dhart/discussions/54
I reverted the primId
part as suggested in #54.
The float[] overloads do come with a significant speed gain (#46) hence my suggestion to expand it to other methods as well.
Thanks for cleaning it up, its much easier to see. I was a bit confused between https://github.com/cadop/dhart/issues/38 which was about changing the c_interface and https://github.com/cadop/dhart/pull/46 about the C# interface.
I'll writeup new performance tests for passing a float vs the new flatten method to make sure that difference is still there (the flatten array method I ended up using as an overload was much faster than the original one). Then I can make pull this
sorry it took so long, i pulled this and made some changes to how its called with the overloads in this commit https://github.com/cadop/dhart/commit/65107be4899bcd4ca32a341733f5bba0aced2429
If you notice it causes any issues please let me know
Can you split up PRs to be just one feature?
I thought this was referring to https://github.com/cadop/dhart/pull/46 (i.e. the float[] overload for c#), but this PR seems like there are a lot of additions to a codebase for accepting a float on a few functions.
It seems like part of this is referring to this issue: https://github.com/cadop/dhart/issues/38 which I set as on hold because its not clear to me how the benefits outweigh the increase in managing the codebase.