It detects as a core2 in gmp-4.2.4, so that agrees with what I found timing wise.
I'm a little surprised that the old magma version on sage.math is beating us at integer multiplication, and from a very early point (WELL before the FFT range). Surely it was linked against a static gmp for amd64 with Pierrick Gaudry's patches. So I don't understand quite how it can beat us. When I try building MPIR with the amd64 patches, it is certainly not better. Bill. 2008/12/22 Bill Hart <goodwillh...@googlemail.com>: > The performance of both the amd64 and core2 patches seems to suck on > this machine. > > Here is a graph of Magma vs FLINT at polynomial multiplication: > > http://sage.math.washington.edu/home/wbhart/flint-trunk/graphing/polymul.png > > The leftmost column of that graph is simply integer multiplication. > That's using the core2 patches. It's worse with amd64. > > It's not even clear to me what this machine should be classed as. > > No hints from gmp-4.2.2. It didn't build for me. I'll try a later version. > > Bill. > > 2008/12/22 Bill Hart <goodwillh...@googlemail.com>: >> >> I just noticed that MPIR uses generic C for all functions on the new >> sage.math. It's has Intel Xeon(R) X7460 CPU's at 2.66 GHz, however >> config.guess says p6???-unknown-linux-gnu. As it doesn't recognise it, >> it builds without assembly support. >> >> 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 -~----------~----~----~----~------~----~------~--~---