Dears, I have see some performance regression with the benchmark through "openssl speed sha1" command (example).
The reason is quite simple, the test methodology has changed in 1.0.1.
(simplified form)
Before:
EVP_MD_CTX_init()
EVP_MD_CTX_set_flags()
loop() {
EVP_DigestInit_ex()
EVP_DigestUpdate()
EVP_DigestFinal_ex()
}
EVP_MD_CTX_cleanup()
After:
loop() {
EVP_MD_CTX_init()
EVP_MD_CTX_set_flags()
EVP_DigestInit_ex()
EVP_DigestUpdate()
EVP_DigestFinal_ex()
EVP_MD_CTX_cleanup()
}
Can it be possible to rool back on the previous algorithm to have comparable
result ?
Note: The performance lost is about 20% in my case.
Regards
--
Alexandre Martins
NETASQ -- We secure IT
smime.p7s
Description: S/MIME cryptographic signature
