Is this library Vc 2.0?

[bernhardmgruber]

Hi! Your previous library Vc promises that after successful standardization of std::simd, development of Vc will be rebased onto std::simd under the name Vc 2.0. However, I have been watching vir-simd for a while now and started to wonder whether this project is starting to become Vc 2.0. So I have some questions:

• Is vir-simd just a prototyping playground for proposals around std::simd or is intended as a solid implementation of std::simd for people stuck on older compilers?
• Would you advice against using vir-simd in production as a replacement for Vc 1.4? Some people at CERN are looking for an upgrade path.
• How does the std-simd repository compare to vir-simd?

Depending on your answers I may want to advertise vir-simd on the Vc repository. Thx!

[mattkretz]

This is not what I had in mind for Vc 2.0 but rather born out of necessity for those who want to use std::experimental::simd. After all std::simd still has not landed 😢 and my plate is more than too full with moving that along.

• My prototype for std::simd is in https://github.com/mattkretz/simd-prototyping. This here is not going to be std::simd.

• At this point I think it's better to move from Vc to vir-simd / stdx::simd and move again when std::simd can be used. I do intend to support both the libstdc++ simd implementation and this repo for as long as I can (which is not much of a promise, of course). But I've also requested to increase the bus-factor of this internally at GSI (nothing reliable either, sorry).

• The std-simd repo is where I developed what became the libstdc++ TS simd implementation. It is outdated as development moved to the GCC repo.

PS: Oh, and I'm really sorry for the too late answer. 😞