https://gcc.gnu.org/bugzilla/show_bug.cgi?id=52473
--- Comment #16 from Dominique d'Humieres <dominiq at lps dot ens.fr> --- For the original test compiled with -Ofast, I get Testing explicit DO loops Dim = 1 Elapsed CPU time = 0.846953988 Dim = 2 Elapsed CPU time = 0.724469006 Dim = 3 Elapsed CPU time = 0.680019855 Testing built-in cshift Dim = 1 Elapsed CPU time = 0.506424904 Dim = 2 Elapsed CPU time = 0.391590118 Dim = 3 Elapsed CPU time = 0.352367163 An for the IDRIS test: ===================== Call to test_eoshift ===================== Order of matrix: 1000 test_eoshift> 1) EOSHIFT Used CPU time ==> 2.976 ms test__eoshift> 2) DO loop Used CPU time ==> 0.832 ms Results OK ==================== Call to test_cshift ==================== Order of matrix: 1000 test_cshift> 1) CSHIFT Used CPU time ==> 1.377 ms test__cshift> 2) DO loop Used CPU time ==> 1.051 ms Results OK ===================== Call to test_reshape ===================== Order of matrix: 1000 test__reshape> 1) RESHAPE Used CPU time ==> 4.909 ms test__reshape> 2) DO loop Used CPU time ==> 0.877 ms Results OK May be this PR could be closed.