Re: [PATCH V4] Remove empty loop with assumed finiteness (PR tree-optimization/89713)

2019-06-04 Thread Marc Glisse
On Tue, 4 Jun 2019, Feng Xue OS wrote:  I think we should turn this option on by default, document that and note  that some languages (C++) say loops terminate. To enable this option at -O2 is not very suitable, seems to be more aggressive. Better to turn it on at -O3. Why wouldn't it be su

[PATCH V4] Remove empty loop with assumed finiteness (PR tree-optimization/89713)

2019-06-03 Thread Feng Xue OS
> I think we should turn this option on by default, document that and note > that some languages (C++) say loops terminate. To enable this option at -O2 is not very suitable, seems to be more aggressive. Better to turn it on at -O3. >> +   /* Avoid doing so for OpenACC abstraction calls >> +