Nick Glover wrote: > Maybe the P-1 bounds calculation accounts for the slightly slower > than normal iteration time that 8907359 would have on a P4 because > of the roundoff checking (since it is very close to the P4 512K > FFT limit).
I doubt it. As I explained last July 17 (Mersenne Digest #981), the Prime95 P-1 limit-choosing algorithm didn't consider iteration times (or roundoff checking). >From my posting then: "The algorithm is independent of specific systems or CPUs. That is, it does not consider or estimate actual time needed to perform an operation, or consider a given CPU's relative performance on integer vs. floating-point instructions. Its basic unit of measure for the cost of a procedure is the FFT "squaring" (= a transform, a convolution, then another transform). For GCDs, which do FFT multiplications that are not squarings, the algorithm estimates the number of FFT transforms needed by the GCD, then divides by two to get the equivalent number of squarings." However, any difference in FFT size between a P4 and other CPU, because of SSE support/nonsupport, could make a difference to the algorithm because it _does_ take FFT size into account. My examination was of V22 source code. I've not yet seen V23 source code -- there might be some change in the V23 algorithm. Richard Woods _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
