Wolf Vollprecht
Wolf Vollprecht
Good to hear. If you want to contribute and open a PR that implements this for the free functions, that would be very appreciated :)
Note I am using uvector here because std::vector uses a bitmask ... :)
So far we only have an implementation of the std::reduce interface (https://github.com/QuantStack/xsimd/blob/master/include/xsimd/stl/algorithms.hpp#L144) Furthermore, we implemented some of the ideas you mention in xtensor (https://github.com/QuantStack/xtensor) where we have a general framework...
are you cross-compiling? If not, `-march=native` should probably also work.
This might now be a problem from our side, if the type or type conversion is not correctly implemented... Can you try with only integer for example?
hey, cool, that's good to hear. we should keep this issue open to investigate in the future. it might be related to 64-bit integers and the NEON version.
Would you be interested in providing a PR for this? That would be awesome! Btw. thanks for filing all these bugs, it's very helpful!
I'm very open to suggestions regarding the interface.
This is a fast implementation of filling a vector with a sequence of numbers. Like numpy arange
probably making a free function like `xsimd::sadd(a, b)` would be the way to go. Then you could implement it for the different architectures in the kernels as a xsimd::kernel::sadd(..). And...