Skip to content

v0.16.0

Choose a tag to compare

@tannevaled tannevaled released this 05 Oct 16:23
· 21 commits to main since this release
647224a

arm64: single-precision vector float arithmetic on the .4S arrangement (#19). VFADD4S, VFSUB4S, VFMUL4S, VFMLA4S, VFMLS4S and VFNEG4S are emitted as mnemonics, which Go 1.27's assembler encodes, and checked against golden encodings. They are additive, next to the existing .2D helpers. They are used by go-fft's float32 NEON kernels.