https://go.dev/cl/626277 (in 1.24) added SIMD implementations for swissmap (still 8-byte) control word operations for amd64. We could do similar for other architectures, if the SIMD implementations improve performance. ARM64 is the obvious target here given its wide use.
https://go.dev/cl/626277 (in 1.24) added SIMD implementations for swissmap (still 8-byte) control word operations for amd64. We could do similar for other architectures, if the SIMD implementations improve performance. ARM64 is the obvious target here given its wide use.