I'd like to draw attention to David Harvey's new middle product algorithm, which was announced on the GMP development list:
http://gmplib.org/list-archives/gmp-devel/2009-June/thread.html His substantial paper on the subject is linked from his website: http://cims.nyu.edu/~harvey/ Notice also the code available on his website, the majority of which we can use. The one part I see on a cursory examination, which we cannot use is the midmul_basecase.asm, but it looks like it relies on mul_2 and addmul_2, which we already have tuned relatively optimally. Jason Moxham, are you interested at all in putting together a mulmid_basecase.as for K8 that we can use, given the mpn_mulmid_basecase.c implementation that is already there? Cycles per limb are listed in David's paper? The main important application for us would be speeding up division - by a considerable factor it seems. Bill. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en -~----------~----~----~----~------~----~------~--~---