Open The-Ludwig opened 11 months ago
But I'd suggest dropping the Eigen dependency. In C++ it's easy to work on raw buffers and we could pass pointers to Eigen buffers to functions which only use standard library data structures, like here
Yes, this is a great idea.
Before reading the second comment I already wanted to say that we should use your implementation without Eigen. Also in your implementation the output is not reversed.
However, we need to have a thorough documentation of the boundary conditions. The ones we used are not obvious from the recursive pattern. Probably the notes we have gathered can help with that.
Amazing! Let's make it happen when we find the time.
I know of at least one large neutrino detector which uses a custom implementation in their analysis code.
We should at some point compare the speed to DeBoors fortran implementation. Am 1. Dez. 2023, um 17:33, Jan Ullmann @.**@.>> schrieb:
Yes, this is a great idea.
Before reading the second comment I already wanted to say that we should use your implementation without Eigen. Also in your implementation the output is not reversed.
However, we need to have a thorough documentation of the boundary conditions. The ones we used are not obvious from the recursive pattern. Probably the notes we have gathered can help with that.
— Reply to this email directly, view it on GitHubhttps://github.com/ullmannJan/B-Splines/issues/2#issuecomment-1836426374, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AHRUCHGRPD74R6FE22ZNLKLYHIBELAVCNFSM6AAAAABADA77FGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZWGQZDMMZXGQ. You are receiving this because you authored the thread.Message ID: @.***>
Nice! I remember this. And actually I think this might be useful for some. I don't know if there really is a standard implementation in C++ for BSplines.
What do you think about publishing this to CONAN? Then people could easily use it.