Some benchmarks of various SHA1 implementations at various input lengths
Building and running
Build via:
make
Run via:
./bench-openssl
./bench-git
./bench-nayuki
Result notes:
All 3 implementations have similar performance at the tested input sizes 1 bytes -> 512 bytes
All 3 implementations exhibit performance drops at increasing input block size multiples (512 bits / 64 bytes), though a bit surprisingly not exactly at the boundary but 9 bytes prior:
- (1*64) - 9 = 55
- (2*64) - 9 = 119
- (3*64) - 9 = 183
- (4*64) - 9 = 247
As visualized here: