> > The little work I did on the wrapper did build correctly on my x86 machine. > > didier >
FYI, if that wasn't clear before, I'm taking over this project :) . Yozo Hida just released another version of qd and an spkg is available here: http://sage.math.washington.edu/home/dfdeshom/custom/sage-2.0/spkg/optional/quaddouble-2.2.spkg This one should build without any problems on sage.math There is a C interface to the C++ code in the qd package (for arithmetic operations) but my benchmarks show that the C interface actually outperforms the original C++ code in pyrex. For example: C++: {{{ Loading SAGE library. Current Mercurial branch is: qd-cpp sage: w = QuadDoubleElement (RR(234.4/random())) ; r= QuadDoubleElement (RR(345.434*random()));w;r +4.367546975E2 +9.213433155E0 sage: time [r*w for i in range(5*10^5)]; CPU times: user 1.96 s, sys: 0.14 s, total: 2.10 s Wall time: 2.15 }}} vs C interface: {{{ Loading SAGE library. Current Mercurial branch is: c-qd sage: w = QuadDoubleElement (RR(234.4/random())) ; r= QuadDoubleElement (RR(345.434*random()));w;r +3.169329699E2 +9.460401581E1 sage: time [r*w for i in range(5*10^5)]; CPU times: user 1.72 s, sys: 0.22 s, total: 1.93 s Wall time: 1.98 }}} I'm guessing operator overloading in C++ is to blame, although it just might be my inexperience with pyrex/C++. Here is a patch that implements quad double numbers using the C interface: http://sage.math.washington.edu/home/dfdeshom/custom/patches/qdrf-c-interface.txt didier --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
