There are major failures in the lapack 3.1.1 tests with gcc 2.95.3 and in the full range up to gcc 4.2.1. The problem seems to appear on 32-bit i386, when the routines chgeqz and zhgeqz are compiled with -O2. With 2.95.3 I get: cgd.out: Matrix order= 8, type=15, seed=2866, 214, 1,1861, result 5 is 8.389E+06 Matrix order= 10, type=15, seed= 362, 860,1065,1109, result 5 is 8.389E+06 Matrix order= 12, type=15, seed=2619,3620,1934,2301, result 5 is 8.389E+06 Matrix order= 20, type=15, seed=1185,3266, 461, 701, result 5 is 8.389E+06 CGV drivers: 4 out of 1092 tests failed to pass the threshold zgd.out: Matrix order= 8, type=15, seed=2866, 214, 1,1861, result 5 is 4.504E+15 Matrix order= 10, type=15, seed= 362, 860,1065,1109, result 5 is 4.504E+15 Matrix order= 12, type=15, seed=2619,3620,1934,2301, result 5 is 4.504E+15 Matrix order= 20, type=15, seed=1185,3266, 461, 701, result 5 is 4.504E+15 ZGV drivers: 4 out of 1092 tests failed to pass the threshold (notice that the failing cases are the same) With 4.2.1: CGV drivers: 65 out of 1092 tests failed to pass the threshold ZGV drivers: 66 out of 1092 tests failed to pass the threshold This may be similar to http://gcc.gnu.org/bugzilla/show_bug.cgi?id=1645
*** This bug has been marked as a duplicate of 323 ***