We define the number of bytes to be the number of *input* bytes. We define a 'char' to be a code point (between 1 and 4 bytes). =========================== Using ICU version 72.1 testcases: 4 input detected as UTF16 little-endian current system detected as haswell =========================== convert_utf32_to_latin1+haswell, input size: 328672, iterations: 30000, dataset: unicode_lipsum/wikipedia_mars/esperanto.utflatin32.txt 0.149 ins/byte, 0.051 cycle/byte, 88.806 GB/s (38.6 %), 4.504 GHz, 2.944 ins/cycle 0.597 ins/char, 0.203 cycle/char, 22.202 Gc/s (38.6 %) 4.00 byte/char 3701.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== convert_utf32_to_latin1+haswell, input size: 1729220, iterations: 30000, dataset: unicode_lipsum/wikipedia_mars/french.utflatin32.txt 0.149 ins/byte, 0.048 cycle/byte, 92.388 GB/s (43.9 %), 4.430 GHz, 3.099 ins/cycle 0.594 ins/char, 0.192 cycle/char, 23.097 Gc/s (43.9 %) 4.00 byte/char 18717.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== convert_utf32_to_latin1+haswell, input size: 797324, iterations: 30000, dataset: unicode_lipsum/wikipedia_mars/german.utflatin32.txt 0.149 ins/byte, 0.050 cycle/byte, 79.836 GB/s (49.2 %), 4.020 GHz, 2.953 ins/cycle 0.595 ins/char, 0.201 cycle/char, 19.959 Gc/s (49.2 %) 4.00 byte/char 9987.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== convert_utf32_to_latin1+haswell, input size: 1086972, iterations: 30000, dataset: unicode_lipsum/wikipedia_mars/portuguese.utflatin32.txt 0.149 ins/byte, 0.049 cycle/byte, 92.642 GB/s (17.6 %), 4.569 GHz, 3.016 ins/cycle 0.595 ins/char, 0.197 cycle/char, 23.161 Gc/s (17.6 %) 4.00 byte/char 11733.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I).