Skip to content

Modified of Agner Fog's vectorclass to use SIMDE to support non x86 platforms

License

Notifications You must be signed in to change notification settings

Remnant44/simde_vectorclass

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

95 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Vectorclass SIMDE (VCLSimde)

This is a fork of Agner Fog's excellent vectorclass library, modified to use SIMDe as an intermediary for emulation of SIMD instruction sets that may not be present.

Explicit goals:

  • Use Vec8f or Vec16f as general purpose abstract SIMD blocks
  • Zero-cost abstraction when writing to the native ISA
  • Transparent fallback to AVX2/SSE as needed
  • Transparent use of native intrinsics where possible
  • High performance conversion to WebAssembly and NEON

Explicit non-goals:

  • Support of narrower than 128bit SIMD

The original vectorclass supports the first three goals efficiently. Targeting SIMDe as an intermediary allows for the last two.

The original VCL is HERE

About

Modified of Agner Fog's vectorclass to use SIMDE to support non x86 platforms

Resources

License

Stars

Watchers

Forks

Languages

  • C++ 99.3%
  • C 0.7%