It turns out that here Apple is not doing anything wrong. (after having a discussion here: https://discussions.apple.com/message/17795537)
It's improper use of BLAS by CVXOPT people that triggers errors. The C code in question violates BLAS convention that incx=0 in these kinds of BLAS functions is illegal. Please see p.30, 1st paragraph of Sect. 2.5.3. http://www.netlib.org/blas/blast-forum/chapter2.pdf There is also a description of xAXPBY function (which is a slight extension of xAXBY which you presume to be buggy) on p.47 which says that if incx (or incy) is set to 0 then an error flag must be raised. So this is a CVXOPT bug, as far as I can see, Apple BLAS is not guilty here. So it seems that we can proceed without Atlas on OSX 10.7, after all... Oh well. Dima On Tuesday, 6 March 2012 18:30:19 UTC+8, Dima Pasechnik wrote: > > OK, guys, bad news, we will have to build Altlas on Lion! > Or wait for Apple to fix their blas... > I checked the attached C code on sqrt5, both with gcc and with cc (i.e. > clang), and it's all > confirmed. > > The correct output is: > nash:tmp dima$ ./test_blas > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [-0.50 0.50 1.50 2.50 3.50 4.50 5.50 6.50 ] > y = > [-1.00 0.00 1.00 2.00 3.00 4.00 5.00 ] > nash:tmp dima$ ./test_cblas > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [-0.50 0.50 1.50 2.50 3.50 4.50 5.50 6.50 ] > y = > [-1.00 0.00 1.00 2.00 3.00 4.00 5.00 ] > > And on sqrt5 one gets: > > sqrt5:tmp dima$ ./test_blas > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [-0.50 0.50 1.50 2.50 3.50 4.50 5.50 ] > sqrt5:tmp dima$ ./test_cblas > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [ 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 ] > y = > [-0.50 0.50 1.50 2.50 3.50 4.50 5.50 ] > > > > -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
