http://llvm.org/bugs/show_bug.cgi?id=11771
Erik Schnetter <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|INVALID | --- Comment #7 from Erik Schnetter <[email protected]> 2012-01-18 10:39:15 CST --- I modified my test case to avoid this undefined behaviour (to the best of my knowledge). Unfortunately, this slightly shifts the problem in my overall code. I attach a new self-contained code that shows the failure. It should be silent if all is fine; however, when built with $ clang -o test_2fabs test_2fabs.cl using $ clang -v clang version 3.1 (trunk 148238) Target: x86_64-apple-darwin11.2.0 Thread model: posix this code outputs $ ./test_2fabs FAIL: signbit type=float2 val=-0 res=-1 good=-1 FAIL: signbit type=float2 val=-0.10000000149011612 res=0 good=-1 FAIL: signbit type=float2 val=-0.10000000149011612 res=-1 good=-1 FAIL: signbit type=float2 val=-0 res=0 good=-1 indicating that the "signbit2" implementation contained in this code near the top fails to correctly extract the sign bit for the second vector component. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. _______________________________________________ LLVMbugs mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
