Re: [Cfrg] Re: universal MACs

[Ken Raeburn <raeburn at MIT.EDU>]

On Jan 17, 2005, at 10:17, David A. McGrew wrote:
> For others on the list: anyone else interested in implementing the 
> MAC?  It would be great to hear a 'second' on the test vectors, and it 
> would be interesting to hear about implementation experience in other 
> environments.  In particular, it will be useful to hear how the hash 
> performs on platforms which lack fast floating-point multiplies, but 
> which have a decent integer-based multiply (see Section 5 of "The 
> Poly1305-AES message-authentication code").

I got the same results from a Python implementation 
(www.mit.edu/~raeburn/poly1305aes.py) for the test cases in appendix A 
of the paper.

It's using the generic Python large-integer support, totally untuned 
for this specific application, and openssl's AES implementation via a 
small C module.  I doubt the performance numbers for this setup are 
going to be anything interesting. :-)

Ken


_______________________________________________
Cfrg mailing list
Cfrg at ietf.org
https://www1.ietf.org/mailman/listinfo/cfrg