Arnaud Gardelein <agardel...@yahoo.fr> writes: > Solved by updating from libatlas3gf-base 3.8.3-27 to 3.8.3-28. Thanks.
I also tried this route. Upgrading to 3.8.3-29 delayed the crash quite a bit, but still crashed eventually: Program terminated with signal 4, Illegal instruction. #0 0xb1be8583 in ATL_zgemvC_a1_x1_b0_y1 () from /usr/lib/libblas.so.3gf (gdb) x/i $eip 0xb1be8583 <ATL_zgemvC_a1_x1_b0_y1+259>: xorpd %xmm6,%xmm6 $ /usr/sbin/update-alternatives --display libblas.so.3gf libblas.so.3gf - auto mode link currently points to /usr/lib/atlas-base/atlas/libblas.so.3gf /usr/lib/atlas-base/atlas/libblas.so.3gf - priority 35 slave libatlas.so.3gf: /usr/lib/atlas-base/libatlas.so.3gf slave libcblas.so.3gf: /usr/lib/atlas-base/libcblas.so.3gf slave libf77blas.so.3gf: /usr/lib/atlas-base/libf77blas.so.3gf slave liblapack_atlas.so.3gf: /usr/lib/atlas-base/liblapack_atlas.so.3gf /usr/lib/libblas/libblas.so.3gf - priority 10 Current 'best' version is '/usr/lib/atlas-base/atlas/libblas.so.3gf'. Looks like some SSE2 instructions are still left in the generic package. I'm running unpackaged linear algebra tests on an AMD Athlon(tm) XP 1800+ with flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up Please don't consider this issue fixed yet. -- Thanks, Feri. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
