Build Options for maximum performance of SHA256 digest calculation

[opensslmailing]

Hello team,

I  want  to  build  openssl  for  Windows  to  achieve  best possible
performance on my machine (Core i5 6500)  of  SHA256  digest  calculation for large files in a C++
program. What options do I have to use to achieve this?

P.S. Currently my program calculates  about 330 Mibyte/s wheeas other
programs  using  OpenSSL reach about 400 MiByte/s on the same machine
with the same file

Regards
Matthias