Repository navigation
v0.16.0
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.