Implement SSE code for ARM processors for best performance (similar to GroupSse2Impl).

greg7mdp / parallel-hashmap

A family of header-only, very fast and memory-friendly hashmap and btree containers.

https://greg7mdp.github.io/parallel-hashmap/

Apache License 2.0

2.52k stars 238 forks source link

Implement SSE code for ARM processors for best performance (similar to GroupSse2Impl). #27

Closed devshgraphicsprogramming closed 2 years ago

devshgraphicsprogramming commented 5 years ago

I wonder what happens to your SSE code on ARM CPUs that only have NEON.

greg7mdp commented 5 years ago

If SSE is not available, the code defaults to a non-SSE, portable C++ implementation (see GroupPortableImpl). It won't be quite as fast but it will work.

greg7mdp commented 5 years ago

We could probably add ARM support by using https://github.com/jratcliff63367/sse2neon

devshgraphicsprogramming commented 5 years ago

ok, you might want to open an issue for this, my team could contribute at some point in 2020

greg7mdp commented 5 years ago

Sure, I have reopened the issue.

greg7mdp commented 2 years ago

see https://github.com/abseil/abseil-cpp/issues/1096