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: 18 input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1699376, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/arabic.utf32.txt 7.501 ins/byte, 1.411 cycle/byte, 3.190 GB/s (4.6 %), 4.502 GHz, 5.315 ins/cycle 7.501 ins/char, 1.411 cycle/char, 3.190 Gc/s (4.6 %) 1.00 byte/char 133187.0 ns utf8_length_from_utf32+haswell, input size: 1699376, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/arabic.utf32.txt 1.376 ins/byte, 0.288 cycle/byte, 15.638 GB/s (6.6 %), 4.505 GHz, 4.776 ins/cycle 1.376 ins/char, 0.288 cycle/char, 15.638 Gc/s (6.6 %) 1.00 byte/char 27168.0 ns utf8_length_from_utf32+icelake, input size: 1699376, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/arabic.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1699376, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/arabic.utf32.txt 3.251 ins/byte, 0.568 cycle/byte, 7.879 GB/s (3.7 %), 4.473 GHz, 5.726 ins/cycle 3.251 ins/char, 0.568 cycle/char, 7.879 Gc/s (3.7 %) 1.00 byte/char 53918.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 548832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/chinese.utf32.txt 7.502 ins/byte, 1.414 cycle/byte, 3.179 GB/s (3.8 %), 4.495 GHz, 5.305 ins/cycle 7.502 ins/char, 1.414 cycle/char, 3.179 Gc/s (3.8 %) 1.00 byte/char 43162.0 ns utf8_length_from_utf32+haswell, input size: 548832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/chinese.utf32.txt 1.377 ins/byte, 0.291 cycle/byte, 15.349 GB/s (15.2 %), 4.464 GHz, 4.736 ins/cycle 1.377 ins/char, 0.291 cycle/char, 15.349 Gc/s (15.2 %) 1.00 byte/char 8939.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 548832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/chinese.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 548832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/chinese.utf32.txt 3.252 ins/byte, 0.568 cycle/byte, 7.644 GB/s (4.6 %), 4.343 GHz, 5.724 ins/cycle 3.252 ins/char, 0.568 cycle/char, 7.644 Gc/s (4.6 %) 1.00 byte/char 17950.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 575328, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/czech.utf32.txt 7.502 ins/byte, 1.415 cycle/byte, 3.178 GB/s (4.2 %), 4.498 GHz, 5.301 ins/cycle 7.502 ins/char, 1.415 cycle/char, 3.178 Gc/s (4.2 %) 1.00 byte/char 45257.0 ns utf8_length_from_utf32+haswell, input size: 575328, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/czech.utf32.txt 1.377 ins/byte, 0.288 cycle/byte, 15.721 GB/s (19.1 %), 4.531 GHz, 4.778 ins/cycle 1.377 ins/char, 0.288 cycle/char, 15.721 Gc/s (19.1 %) 1.00 byte/char 9149.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 575328, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/czech.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 575328, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/czech.utf32.txt 3.252 ins/byte, 0.572 cycle/byte, 7.801 GB/s (7.6 %), 4.464 GHz, 5.683 ins/cycle 3.252 ins/char, 0.572 cycle/char, 7.801 Gc/s (7.6 %) 1.00 byte/char 18438.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1550036, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/english.utf32.txt 7.501 ins/byte, 1.411 cycle/byte, 3.189 GB/s (4.5 %), 4.500 GHz, 5.315 ins/cycle 7.501 ins/char, 1.411 cycle/char, 3.189 Gc/s (4.5 %) 1.00 byte/char 121523.0 ns utf8_length_from_utf32+haswell, input size: 1550036, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/english.utf32.txt 1.376 ins/byte, 0.289 cycle/byte, 15.585 GB/s (7.7 %), 4.500 GHz, 4.765 ins/cycle 1.376 ins/char, 0.289 cycle/char, 15.585 Gc/s (7.7 %) 1.00 byte/char 24864.0 ns utf8_length_from_utf32+icelake, input size: 1550036, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/english.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1550036, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/english.utf32.txt 3.251 ins/byte, 0.568 cycle/byte, 7.837 GB/s (3.1 %), 4.453 GHz, 5.721 ins/cycle 3.251 ins/char, 0.568 cycle/char, 7.837 Gc/s (3.1 %) 1.00 byte/char 49447.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 336500, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/esperanto.utf32.txt 7.503 ins/byte, 1.411 cycle/byte, 3.212 GB/s (6.4 %), 4.532 GHz, 5.317 ins/cycle 7.503 ins/char, 1.411 cycle/char, 3.212 Gc/s (6.4 %) 1.00 byte/char 26190.0 ns utf8_length_from_utf32+haswell, input size: 336500, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/esperanto.utf32.txt 1.379 ins/byte, 0.288 cycle/byte, 14.689 GB/s (18.2 %), 4.234 GHz, 4.786 ins/cycle 1.379 ins/char, 0.288 cycle/char, 14.689 Gc/s (18.2 %) 1.00 byte/char 5727.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 336500, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/esperanto.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 336500, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/esperanto.utf32.txt 3.254 ins/byte, 0.569 cycle/byte, 7.924 GB/s (13.1 %), 4.513 GHz, 5.714 ins/cycle 3.254 ins/char, 0.569 cycle/char, 7.924 Gc/s (13.1 %) 1.00 byte/char 10616.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1739468, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/french.utf32.txt 7.501 ins/byte, 1.413 cycle/byte, 3.200 GB/s (4.5 %), 4.522 GHz, 5.307 ins/cycle 7.501 ins/char, 1.413 cycle/char, 3.200 Gc/s (4.5 %) 1.00 byte/char 135911.0 ns utf8_length_from_utf32+haswell, input size: 1739468, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/french.utf32.txt 1.376 ins/byte, 0.289 cycle/byte, 15.724 GB/s (8.7 %), 4.543 GHz, 4.762 ins/cycle 1.376 ins/char, 0.289 cycle/char, 15.724 Gc/s (8.7 %) 1.00 byte/char 27657.0 ns utf8_length_from_utf32+icelake, input size: 1739468, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/french.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1739468, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/french.utf32.txt 3.251 ins/byte, 0.569 cycle/byte, 7.803 GB/s (12.0 %), 4.437 GHz, 5.717 ins/cycle 3.251 ins/char, 0.569 cycle/char, 7.803 Gc/s (12.0 %) 1.00 byte/char 55734.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 804860, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/german.utf32.txt 7.501 ins/byte, 1.408 cycle/byte, 3.173 GB/s (3.0 %), 4.467 GHz, 5.328 ins/cycle 7.501 ins/char, 1.408 cycle/char, 3.173 Gc/s (3.0 %) 1.00 byte/char 63415.0 ns utf8_length_from_utf32+haswell, input size: 804860, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/german.utf32.txt 1.377 ins/byte, 0.293 cycle/byte, 15.244 GB/s (13.2 %), 4.459 GHz, 4.707 ins/cycle 1.377 ins/char, 0.293 cycle/char, 15.244 Gc/s (13.2 %) 1.00 byte/char 13200.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 804860, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/german.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 804860, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/german.utf32.txt 3.252 ins/byte, 0.569 cycle/byte, 7.937 GB/s (6.7 %), 4.520 GHz, 5.710 ins/cycle 3.252 ins/char, 0.569 cycle/char, 7.937 Gc/s (6.7 %) 1.00 byte/char 25352.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 571996, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/greek.utf32.txt 7.502 ins/byte, 1.411 cycle/byte, 3.155 GB/s (3.0 %), 4.451 GHz, 5.317 ins/cycle 7.502 ins/char, 1.411 cycle/char, 3.155 Gc/s (3.0 %) 1.00 byte/char 45326.0 ns utf8_length_from_utf32+haswell, input size: 571996, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/greek.utf32.txt 1.378 ins/byte, 0.290 cycle/byte, 15.511 GB/s (20.9 %), 4.491 GHz, 4.759 ins/cycle 1.378 ins/char, 0.290 cycle/char, 15.511 Gc/s (20.9 %) 1.00 byte/char 9219.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 571996, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/greek.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 571996, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/greek.utf32.txt 3.252 ins/byte, 0.569 cycle/byte, 7.815 GB/s (15.3 %), 4.444 GHz, 5.719 ins/cycle 3.252 ins/char, 0.569 cycle/char, 7.815 Gc/s (15.3 %) 1.00 byte/char 18298.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 585404, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hebrew.utf32.txt 7.502 ins/byte, 1.426 cycle/byte, 3.165 GB/s (3.9 %), 4.514 GHz, 5.261 ins/cycle 7.502 ins/char, 1.426 cycle/char, 3.165 Gc/s (3.9 %) 1.00 byte/char 46235.0 ns utf8_length_from_utf32+haswell, input size: 585404, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hebrew.utf32.txt 1.378 ins/byte, 0.288 cycle/byte, 15.409 GB/s (18.4 %), 4.433 GHz, 4.788 ins/cycle 1.378 ins/char, 0.288 cycle/char, 15.409 Gc/s (18.4 %) 1.00 byte/char 9498.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 585404, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hebrew.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 585404, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hebrew.utf32.txt 3.252 ins/byte, 0.570 cycle/byte, 7.819 GB/s (9.7 %), 4.459 GHz, 5.703 ins/cycle 3.252 ins/char, 0.570 cycle/char, 7.819 Gc/s (9.7 %) 1.00 byte/char 18718.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1095832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hindi.utf32.txt 7.501 ins/byte, 1.417 cycle/byte, 3.184 GB/s (2.9 %), 4.512 GHz, 5.294 ins/cycle 7.501 ins/char, 1.417 cycle/char, 3.184 Gc/s (2.9 %) 1.00 byte/char 86044.0 ns utf8_length_from_utf32+haswell, input size: 1095832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hindi.utf32.txt 1.376 ins/byte, 0.290 cycle/byte, 15.030 GB/s (6.4 %), 4.361 GHz, 4.743 ins/cycle 1.376 ins/char, 0.290 cycle/char, 15.030 Gc/s (6.4 %) 1.00 byte/char 18228.0 ns utf8_length_from_utf32+icelake, input size: 1095832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hindi.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1095832, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/hindi.utf32.txt 3.251 ins/byte, 0.570 cycle/byte, 7.783 GB/s (4.5 %), 4.434 GHz, 5.707 ins/cycle 3.251 ins/char, 0.570 cycle/char, 7.783 Gc/s (4.5 %) 1.00 byte/char 35200.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 475564, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/japanese.utf32.txt 7.502 ins/byte, 1.417 cycle/byte, 3.158 GB/s (4.7 %), 4.474 GHz, 5.296 ins/cycle 7.502 ins/char, 1.417 cycle/char, 3.158 Gc/s (4.7 %) 1.00 byte/char 37644.0 ns utf8_length_from_utf32+haswell, input size: 475564, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/japanese.utf32.txt 1.378 ins/byte, 0.289 cycle/byte, 15.762 GB/s (20.9 %), 4.554 GHz, 4.769 ins/cycle 1.378 ins/char, 0.289 cycle/char, 15.762 Gc/s (20.9 %) 1.00 byte/char 7543.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 475564, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/japanese.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 475564, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/japanese.utf32.txt 3.253 ins/byte, 0.570 cycle/byte, 7.955 GB/s (14.0 %), 4.534 GHz, 5.707 ins/cycle 3.253 ins/char, 0.570 cycle/char, 7.955 Gc/s (14.0 %) 1.00 byte/char 14946.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 291672, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/korean.utf32.txt 7.503 ins/byte, 1.411 cycle/byte, 3.183 GB/s (6.3 %), 4.492 GHz, 5.316 ins/cycle 7.503 ins/char, 1.411 cycle/char, 3.183 Gc/s (6.3 %) 1.00 byte/char 22908.0 ns utf8_length_from_utf32+haswell, input size: 291672, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/korean.utf32.txt 1.380 ins/byte, 0.290 cycle/byte, 15.131 GB/s (24.5 %), 4.392 GHz, 4.755 ins/cycle 1.380 ins/char, 0.290 cycle/char, 15.131 Gc/s (24.5 %) 1.00 byte/char 4819.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 291672, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/korean.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 291672, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/korean.utf32.txt 3.254 ins/byte, 0.570 cycle/byte, 7.970 GB/s (16.2 %), 4.546 GHz, 5.705 ins/cycle 3.254 ins/char, 0.570 cycle/char, 7.970 Gc/s (16.2 %) 1.00 byte/char 9149.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 498776, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/persan.utf32.txt 7.502 ins/byte, 1.416 cycle/byte, 3.183 GB/s (4.6 %), 4.506 GHz, 5.299 ins/cycle 7.502 ins/char, 1.416 cycle/char, 3.183 Gc/s (4.6 %) 1.00 byte/char 39181.0 ns utf8_length_from_utf32+haswell, input size: 498776, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/persan.utf32.txt 1.378 ins/byte, 0.292 cycle/byte, 14.058 GB/s (5.7 %), 4.108 GHz, 4.716 ins/cycle 1.378 ins/char, 0.292 cycle/char, 14.058 Gc/s (5.7 %) 1.00 byte/char 8870.0 ns utf8_length_from_utf32+icelake, input size: 498776, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/persan.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 498776, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/persan.utf32.txt 3.253 ins/byte, 0.571 cycle/byte, 7.797 GB/s (9.8 %), 4.455 GHz, 5.693 ins/cycle 3.253 ins/char, 0.571 cycle/char, 7.797 Gc/s (9.8 %) 1.00 byte/char 15993.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1094456, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/portuguese.utf32.txt 7.501 ins/byte, 1.411 cycle/byte, 3.177 GB/s (2.6 %), 4.483 GHz, 5.316 ins/cycle 7.501 ins/char, 1.411 cycle/char, 3.177 Gc/s (2.6 %) 1.00 byte/char 86114.0 ns utf8_length_from_utf32+haswell, input size: 1094456, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/portuguese.utf32.txt 1.376 ins/byte, 0.290 cycle/byte, 15.424 GB/s (15.5 %), 4.470 GHz, 4.749 ins/cycle 1.376 ins/char, 0.290 cycle/char, 15.424 Gc/s (15.5 %) 1.00 byte/char 17739.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 1094456, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/portuguese.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1094456, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/portuguese.utf32.txt 3.251 ins/byte, 0.569 cycle/byte, 7.947 GB/s (6.4 %), 4.526 GHz, 5.709 ins/cycle 3.251 ins/char, 0.569 cycle/char, 7.947 Gc/s (6.4 %) 1.00 byte/char 34431.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1248148, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/russian.utf32.txt 7.501 ins/byte, 1.417 cycle/byte, 3.207 GB/s (5.4 %), 4.544 GHz, 5.294 ins/cycle 7.501 ins/char, 1.417 cycle/char, 3.207 Gc/s (5.4 %) 1.00 byte/char 97288.0 ns utf8_length_from_utf32+haswell, input size: 1248148, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/russian.utf32.txt 1.376 ins/byte, 0.291 cycle/byte, 15.406 GB/s (9.6 %), 4.483 GHz, 4.729 ins/cycle 1.376 ins/char, 0.291 cycle/char, 15.406 Gc/s (9.6 %) 1.00 byte/char 20254.0 ns utf8_length_from_utf32+icelake, input size: 1248148, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/russian.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1248148, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/russian.utf32.txt 3.251 ins/byte, 0.568 cycle/byte, 7.797 GB/s (7.0 %), 4.425 GHz, 5.728 ins/cycle 3.251 ins/char, 0.568 cycle/char, 7.797 Gc/s (7.0 %) 1.00 byte/char 40019.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1619032, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/thai.utf32.txt 7.501 ins/byte, 1.410 cycle/byte, 3.191 GB/s (5.2 %), 4.499 GHz, 5.320 ins/cycle 7.501 ins/char, 1.410 cycle/char, 3.191 Gc/s (5.2 %) 1.00 byte/char 126831.0 ns utf8_length_from_utf32+haswell, input size: 1619032, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/thai.utf32.txt 1.376 ins/byte, 0.290 cycle/byte, 15.455 GB/s (7.4 %), 4.479 GHz, 4.748 ins/cycle 1.376 ins/char, 0.290 cycle/char, 15.455 Gc/s (7.4 %) 1.00 byte/char 26190.0 ns utf8_length_from_utf32+icelake, input size: 1619032, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/thai.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1619032, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/thai.utf32.txt 3.251 ins/byte, 0.569 cycle/byte, 7.842 GB/s (3.6 %), 4.460 GHz, 5.716 ins/cycle 3.251 ins/char, 0.569 cycle/char, 7.842 Gc/s (3.6 %) 1.00 byte/char 51612.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 741768, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/turkish.utf32.txt 7.501 ins/byte, 1.413 cycle/byte, 3.191 GB/s (4.2 %), 4.510 GHz, 5.308 ins/cycle 7.501 ins/char, 1.413 cycle/char, 3.191 Gc/s (4.2 %) 1.00 byte/char 58108.0 ns utf8_length_from_utf32+haswell, input size: 741768, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/turkish.utf32.txt 1.377 ins/byte, 0.301 cycle/byte, 15.086 GB/s (18.4 %), 4.541 GHz, 4.573 ins/cycle 1.377 ins/char, 0.301 cycle/char, 15.086 Gc/s (18.4 %) 1.00 byte/char 12292.0 ns WARNING: Measurements are noisy, try increasing iteration count (-I). utf8_length_from_utf32+icelake, input size: 741768, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/turkish.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 741768, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/turkish.utf32.txt 3.251 ins/byte, 0.568 cycle/byte, 7.809 GB/s (6.6 %), 4.438 GHz, 5.721 ins/cycle 3.251 ins/char, 0.568 cycle/char, 7.809 Gc/s (6.6 %) 1.00 byte/char 23746.0 ns input detected as UTF16 little-endian current system detected as haswell =========================== utf8_length_from_utf32+fallback, input size: 1129676, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/vietnamese.utf32.txt 7.501 ins/byte, 1.425 cycle/byte, 3.182 GB/s (3.3 %), 4.534 GHz, 5.263 ins/cycle 7.501 ins/char, 1.425 cycle/char, 3.182 Gc/s (3.3 %) 1.00 byte/char 88767.0 ns utf8_length_from_utf32+haswell, input size: 1129676, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/vietnamese.utf32.txt 1.376 ins/byte, 0.291 cycle/byte, 15.317 GB/s (8.0 %), 4.453 GHz, 4.734 ins/cycle 1.376 ins/char, 0.291 cycle/char, 15.317 Gc/s (8.0 %) 1.00 byte/char 18438.0 ns utf8_length_from_utf32+icelake, input size: 1129676, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/vietnamese.utf32.txt utf8_length_from_utf32+icelake: unsupported by the system utf8_length_from_utf32+westmere, input size: 1129676, iterations: 1000, dataset: unicode_lipsum/wikipedia_mars/vietnamese.utf32.txt 3.251 ins/byte, 0.569 cycle/byte, 7.807 GB/s (4.4 %), 4.445 GHz, 5.710 ins/cycle 3.251 ins/char, 0.569 cycle/char, 7.807 Gc/s (4.4 %) 1.00 byte/char 36177.0 ns