4 ms·
- Byte/Nibble transpose/shuffle for improving compression of binary data (ex. floating point data) - Scalar/SIMD Transpose/Shuffle 8,16,32,64,... bits Dynamic
by powturbo 10y ago
- Byte/Nibble transpose/shuffle for improving compression of binary data (ex. floating point data)
- Scalar/SIMD Transpose/Shuffle 8,16,32,64,... bits
Dynamic CPU detection and JIT scalar/sse/avx2 switching
- 100% C (C++ headers), usage as simple as memcpy
- Ready and simple to use library, no hassless dependencies
+ Byte Transpose
- Fastest byte transpose
+ Nibble Transpose
- nearly as fast as byte transpose
- more efficient in most binary data files, up to 6 times faster than Bitshuffle
- more robust worst case scenario than bitshuffle