On Sunday 20 February 2011 18:19:23 Jason wrote: > On Sunday 20 February 2011 17:24:40 Bill Hart wrote: > > Oooh. That looks good! > > > > Great work Jason. > > > > On 20 February 2011 13:35, Jason <ja...@njkfrudils.plus.com> wrote: > > > On Saturday 19 February 2011 10:52:10 Jason wrote: > > >> On Tuesday 15 February 2011 17:03:24 Jason wrote: > > >> > Hi > > >> > > > >> > A new mpn_mod_1_1 for the nehalem/westmere , previously ran at > > >> > 12.5c/l now runs at 12c/l which is optimal (for the chosen method). > > >> > > > >> > This was running faster on the core2/penryn , but not since I did > > >> > the feed-in > > >> > > > >> > :( > > >> > > > >> > Jason > > >> > > >> New mpn_mod_1_1 for Core2 and another slightly different one for > > >> Penryn. The old code ran at 14.3c/l and the the new code's run at > > >> 13c/l and 13.3c/l > > >> > > >> Jason > > > > > > A new AMD mpn_mod_1_2 used to run at 4.0c/l now runs at 3.5c/l which is > > > optimal. > > > > > > Brian , for the windows conversions note that for all the different > > > cpus , mod_1_1's and mod_1_2 etc are all minor re-arrangements of each > > > other. > > > > > > I've adapted my search program to accept prologue and epilogue code to > > > the loops, this helped find the code above , hopefully it will help the > > > Intel chips as well as they suffer much more from this problem. > > > > > > Jason > > > > > > -- > > > 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. > > A new mpn_mod_1_2 for Core2/Penryn (only one version this time) , used to > run at 7.9c/l now runs at 6.2c/l (6.0c/l is optimal) > > Jason
A new function mpn_not , which is just an inplace mpn_com , for AMD we did run at 1.25c/l now run at 1.0c/l which is optimal on K8 Jason -- 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.