Hi Anton, You're right, any AES mode that allows independent encryption or decryption of blocks can be optimized to perform AES rounds in parallel in the pipeline.
I'm finishing the patches for SHA-2 and polynomial multiplication for GCM and I also intend to provide patches for AES-CTR, AES-CBC, AES-XTS and other modes. Another change in my plans is to provide an optimized version of AES_set_encrypt_key/AES_set_decrypt_key. Regards, Marcelo On Wed, Nov 20, 2013 at 09:43:54AM +1100, Anton Blanchard wrote: > > Hi, > > > Include support for new POWER8 vcipher instructions. It uses > > OPENSSL_ppccap_P to choose which implementation to use, regular or > > POWER8. > > Nice! We should be able to speed up parallelizable AES block modes like > AES-CTR even more. From some test loops I wrote, overlapping 6 > iterations should give us close to another 6x improvement. > > Anton > ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
