Dear Torbjörn, > Without invariance from e.g modexp, I don't believe one can beat > sbpi1_bdiv_r (or the older redc_1). Newer implementation of > sbpi1_bdiv_r make use of the observation that the next quotient can be > computed early, actually almost a whole innerloop invocation early, > which makes its cost very low.
of course I was thinking in the invariant case. I'm looking forward seeing sbpi1_bdiv_r in the public interface, this would be very interesting in GMP-ECM for example. Paul _______________________________________________ gmp-devel mailing list gmp-devel@gmplib.org https://gmplib.org/mailman/listinfo/gmp-devel
