[Bug tree-optimization/78386] Optimization -O2 and higher cause math inconsistency

2016-11-17 Thread gromero at linux dot vnet.ibm.com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78386 Gustavo Romero changed: What|Removed |Added CC||gromero at linux dot vnet.ibm.com ---

[Bug tree-optimization/78386] Optimization -O2 and higher cause math inconsistency

2016-11-17 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78386 --- Comment #6 from Richard Biener --- Should be the default when you don't enable GNU extensions (thus use -std=c11 for example). But use of FMA can also appear without -fexpensive-optimizations if the combiner can synthesize it on RTL (without

[Bug tree-optimization/78386] Optimization -O2 and higher cause math inconsistency

2016-11-16 Thread pinskia at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78386 --- Comment #5 from Andrew Pinski --- Does -ffp-contract=off help?

[Bug tree-optimization/78386] Optimization -O2 and higher cause math inconsistency

2016-11-16 Thread joseph at codesourcery dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78386 --- Comment #4 from joseph at codesourcery dot com --- On Wed, 16 Nov 2016, pinskia at gcc dot gnu.org wrote: > Most likely the use of fma. Which applies with current x86_64 just as it does with powerpc. It should be possible to reduce such a

[Bug tree-optimization/78386] Optimization -O2 and higher cause math inconsistency

2016-11-16 Thread wschmidt at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78386 Bill Schmidt changed: What|Removed |Added Target||powerpc64le-unknown-linux-g