[tahoe-dev] pycryptopp benchmarks on ARM CPU

[Jack Lloyd]

On Tue, Jun 23, 2009 at 08:37:06PM -0600, Zooko O'Whielacronx wrote:

> It is the time it takes to generate a new key pair.  (I suspect that
> the RSA keypair generation in Crypto++ is not optimized.)

To some extent the problem is algorithmic, not implementation:
generating an RSA key pair requires finding 2 large primes, which can
be done relatively quickly using a wide variety of sieves and other
tricks, BUT generating an ECDSA keypair requires simply choosing a
random point and a small number of point multiplications (which are
themselves quite fast (and probably quite well optimized in
Crypto++)).

-Jack