On 4/23/12, Samuel Neves <[email protected]> wrote: > > On big hardware, the fastest SHA-3 candidates (BLAKE, Skein) are very > much closer to MD5 in performance (~5.5 cpb) than SHA-2. Plus, I don't > see any platform where CubeHash16/32 wins over either of them in speed. > > The place where SHA-2 shines is the very low end. Performance there, > however, is usually measured in gates, not cpb. >
The latest performance of Skein and BLAKE that you are mentioning is due the continuous efforts of designers and independent programmers to improve their implementation. As I can see, measurements of SHA-2 are mostly from an OpenSSL implementation that is not as much optimized as the implementations of the 5 SHA-3 finalists. But once SHA-2 is started to be as aggressively optimized as SHA-3 finalists, we will see reports like the one in [1]: "Furthermore, even the fastest finalists will probably offer only a small performance advantage over the current SHA-256 and SHA-512 implementations." Unfortunately, also I do not see any more improvements of the implementations of other SHA-3 candidates that did not enter 2-nd and final round (especially CubeHash, Shabal, BMW, Edon-R, Echo and SIMD). Regards, David Adamson Jr [1] Shay Gueron, Vlad Krasnov, "Parallelizing message schedules to accelerate the computations of hash functions" _______________________________________________ cryptography mailing list [email protected] http://lists.randombit.net/mailman/listinfo/cryptography
