Here are the timings for SSLeay on banana-jr, running the same code as James Washington's BSAFE tests, modified for SSLeay. (This is CBC DES) Encrypting 1 byte at a time: KeyStream Size # KeyStreams Elapsed Time Rate -------------- ------------ ------------ ---- 4194304 1 2.439799 1.639479 2097152 2 2.311199 1.730703 1048576 4 2.753021 1.452949 524288 8 2.572160 1.555113 262144 16 2.578761 1.551132 131072 32 2.627386 1.522426 65536 64 2.554560 1.565827 32768 128 2.510106 1.593558 16384 256 2.691806 1.485991 8192 512 2.381742 1.679443 4096 1024 2.379339 1.681139 2048 2048 2.326281 1.719483 1024 4096 2.526309 1.583338 512 8192 2.470835 1.618886 256 16384 2.305602 1.734905 128 32768 2.368472 1.688852 64 65536 2.324252 1.720984 32 131072 2.514618 1.590699 16 262144 2.391457 1.672620 8 524288 2.552542 1.567065 Encrypting 4 bytes at a time: KeyStream Size # KeyStreams Elapsed Time Rate -------------- ------------ ------------ ---- 4194304 1 2.206622 1.812726 2097152 2 2.050378 1.950860 1048576 4 2.051156 1.950120 524288 8 2.050065 1.951158 262144 16 2.044363 1.956600 131072 32 2.034407 1.966175 65536 64 2.016850 1.983291 32768 128 2.020192 1.980010 16384 256 2.012792 1.987289 8192 512 2.006580 1.993441 4096 1024 1.999419 2.000581 2048 2048 2.009245 1.990798 1024 4096 1.997528 2.002475 512 8192 2.000849 1.999151 256 16384 2.004074 1.995934 128 32768 2.015869 1.984256 64 65536 2.026293 1.974048 32 131072 2.106540 1.898848 16 262144 2.096111 1.908296 8 524288 2.181430 1.833660 I don't know if the machine the BSAFE tests were run on is similar in specs to banana-jr, but if so, SSLeay seems considerably faster. The pattern of numbers agrees with the BSAFE results, and I think the same conclusion can be drawn.