search for: d24805

Hi, Yes, I can see why that would not work. The sinking algorithm in SimplifyCFG isn't particularly clever. In particular it can't reason about memory ordering and aliasing. In unswitch1(), it can identify that the stores correlate because the correlating stores appear in the same relative source order. In unswitch2() they have been permuted, and the algorithm cannot deal with this. This

...2017 at 16:50, Hongbin Zheng <etherzhhb at gmail.com> wrote: > Hi James, > > I have an ad-hoc solution in mind to solve this problem. > But if it can be solved under the framework of GVN"Sink", it is even > better. > > any plan on your https://reviews.llvm.org/D24805? > > Thanks > Hongbin > > > On Wed, Apr 26, 2017 at 2:13 AM, James Molloy <james at jamesmolloy.co.uk> > wrote: > >> Hi, >> >> Yes, I can see why that would not work. >> >> The sinking algorithm in SimplifyCFG isn't particularly clever...

...te: >> >>> Hi James, >>> >>> I have an ad-hoc solution in mind to solve this problem. >>> But if it can be solved under the framework of GVN"Sink", it is even >>> better. >>> >>> any plan on your https://reviews.llvm.org/D24805? >>> >>> Thanks >>> Hongbin >>> >>> >>> On Wed, Apr 26, 2017 at 2:13 AM, James Molloy <james at jamesmolloy.co.uk> >>> wrote: >>> >>>> Hi, >>>> >>>> Yes, I can see why that would not wo...