Dong H. Ahn
Dong H. Ahn
Thinking about this a bit more, I am not sure if the callback is the best mechanism. But what I was thinking is have a way for people like us...
@jprotze: oh nice! It sounds like we already have a baseline to extend :-)
@SumedhArani: by multilevel, I mean an ability to compare two race instances and treat them as "equivalent" or "distinct" with different criteria. One special case would be to reduce those...
@SumedhArani: Yes, let's go with the same leaf point as the criteria to begin with and then try to find ways to program different criteria. It feels like this project...
Hey @SumedhArani: Thank you for the quick response! It is kind of hard to review what has changed without diffs. How about we do this. 1. Fork the vanilla TSan...
I prefer keeping this open until we get a complete solution merged into Archer. I think the current solution need to grow a bit to meet the use case we...
Have you also tried Frontran code that goes into a static library? `-fPIC` on the Fortran code might do but then, this may require `libfortran.a` to be built with `-fPIC`...
Good thoughts! Yes this would be a faster alternative to the current race reduction angle. Two things concern me a bit, none of which is fundamental to the proposal but...
@SumedhArani: I think the main distinction is on-the-fly duplicate detection vs. post mortem detection. As I mentioned above, I think this optimization can be done later once we understand what...
No, that's your topic now :-)