On Tue, 27 Dec 2005 15:17:50 -0500, Dave Roundy wrote:
On Tue, Dec 27, 2005 at 01:36:53AM -0700, David Mercer wrote:
We are pleased to announce Version 0.21a, the initial functioning release of Shaskell, our SHA2 library for (so far) ghc6. It supports sha256 and sha512 hashes in this version, and is "pure": no monads or C libraries are involved!
You state in your abstract that the speed is comparable to a C implementation. Do you have numbers comparing speed with the openssl implementation, or any other implementations?
David Roundy
Doh! I meant to tone down that sentence, and insert the word "hopefully", as I haven't done speed testing in any depth yet, but a trip to the hospital this month left a few loose ends dangling, that being one of them! New abstract with toned down claim inserted to be in the future tense! Thanks for catching that. Performance of the sha512 implementation on 32 bit machines is currently pretty horrid, as would be expected, but the sha256 is much better. Speaking of testing, is anyone aware of any good test harnesses for benchmark comparison testing, it's been quite a while since I cared about such things and am out of touch with that corner of computing! Testing other sha2 libraries for comparison is going to be a bit of a bear, as most of them don't seem to have convenient command line versions/wrappers available, and I loath touching C, except with the traditional "make; make install" ritual of package installation. Thanks, -David Mercer Tucson, AZ