On Dec 2, 2013, at 11:58 AM, Jens Alfke <j...@mooseyard.com> wrote: > > On Dec 2, 2013, at 6:36 AM, ChanMaxthon <xcvi...@me.com> wrote: > >> How about this: treat the data as an array of 64-bit integers (zero-padded) >> and XOR them together? That will be really fast and you will be able to >> asm-optimize it. > > Did you read the original post? That’s almost exactly what he’s already doing > that he wants to improve. > >> Or maybe, use some bits and pieces of AES (namely AES-NI instructions) and >> Merkle-Damgård construction? Since security is no major concern those >> asm-level constructs can be useful. > > Seriously? To quote the last reply I sent, "The OP wants a real-world > solution, not an abstract theoretical example”. > > And guys, didn’t we answer the question already? And didn’t Graham already > say he’d decided on murmurhash? I get the feeling we’re continuing to beat > this dead horse into the ground just to show off what we know about > combinatorics and crypto [myself included].
Amen. I don’t have anything to show off except an old maxim: You can’t make a computer run faster, you can only make it do less work. And the best kind of work is the kind that you don’t have to do. If the file size is identical If the first 8 bytes are identical if MurmurHash(file) is identical if byte-by-byte comparison is identical IT’S THE SAME IMAGE else IT AIN’T else IT AIN’T else IT AIN’T else IT AIN’T There are lots of ways to prove that they are DIFFERENT, and they are fairly quick. Use them first. ---------------------------------------------------------------- Steve Bird Culverson Software - Elegant software that is a pleasure to use. www.Culverson.com (toll free) 1-877-676-8175 _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com