I know there has been some discussion on this list about the theoretical possibility of staying in frequency space and doing a bunch of point-wise squarings in there instead of computing the inverse DWT after every iteration. Now, I'm not suggesting this as a feasible avenue for GIMPS as I've read some of the arguments against it in the archives. I'm just curious as to how it might proceed if you could assume infinite precision, etc. For instance, would it be as simple as:
DWT(x) -> X^n -> IDWT(X^n) == x^n (mod 2^q - 1) (where "n" is now only 2?) Of course I'm simplifying by not including the subtraction by 2, but you get the idea. Also, if it *is* this simple, could there be any modular reduction of the X^n before doing the IDWT? Any explanation would be appreciated. Brendan Younger _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
