When you do this, could you also do: cd tune make tune
and send me the results for each machine. The crossovers have now changed a lot after I fixed numerous performance issues. Bill. On 17 February 2014 21:06, Bill Hart <goodwillh...@googlemail.com> wrote: > Excellent. Here are the (abbreviated) instructions: > > git clone https://github.com/wbhart/mpir > ./configure {check it really detects as a nehalem/atom} > make -j > cd tune > make speed > ./speed -s 1-100 -t 1 mpn_sqr_basecase > > Record the values somewhere. > > Now, on the nehalem, remove the file mpn/x86_64/nehalem/sqr_basecase.as > On the atom, remove mpn/x86_64/atom/sqr_basecase.as > > Now time it again (go to top level directory): > > ./configure > make clean > make -j > cd tune > make speed > ./speed -s 1-100 -t 1 mpn_sqr_basecase > > Check the clock speed reported in both runs is the same on each machine. > > Tell me whether the numbers in the second run are mostly lower or not. > > If they are can you send the numbers to me and the output of > > ./speed -s 1-100 -t 1 mpn_kara_sqr_n > > for each machine. > > Thanks. > > Bill. > > > On 17 February 2014 20:58, Jean-Pierre Flori <jpfl...@gmail.com> wrote: > >> >> >> On Monday, February 17, 2014 7:31:35 PM UTC+1, Bill Hart wrote: >>> >>> I have access to quite a few machines, but currently not atom, nehalem, >>> bobcat, bulldozer or haswell chips. >>> >>> I have access to an atom and a nehalem (Core2). >> > > -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to mpir-devel+unsubscr...@googlegroups.com. To post to this group, send email to mpir-devel@googlegroups.com. Visit this group at http://groups.google.com/group/mpir-devel. For more options, visit https://groups.google.com/groups/opt_out.