Optimization for power

[ghatwala]

Hi @klauspost ,

I have been trying to understand the process for creating optimized go code for power(ppc64le) in here "https://github.com/klauspost/reedsolomon" similar to the optimization done for amd(galois_amd64.go) and arm(galois_arm64.go).

My understanding is below for example on "amd64"
The routines that have been optimized for amd64 and are present in " galois_amd64.go" in klauspost repo.
o galMulSlice
o galMulSliceXor
o sliceXor
The un-optimized implementation of the above routines is in "galois_noasm.go " and this would be the one that is getting invoked for Power currently. ?

To implement the equivalent power assembly - which intel equivalent needs to be considered - AVX2 or SSSE3 ?

[Frank Wessels]

This has been fixed with #88

[Klaus Post]

Indeed!