sebpop wrote:

> Generally assume best intentions.

I agree, and I will keep this in mind.

> These issues aren't caught by assertions and can silently lead to incorrect 
> transformations.

For GCC, I adapted the Omega test (from Bill Pugh) to produce the same 
representation as the classic dependence tests.  And we got more confidence 
that the dependence analysis implementation was correct. I removed the extra 
checks 
https://github.com/gcc-mirror/gcc/commit/49b8fe6c1a585edfeb5dd0f292e05a167f475f68
 10 years down the road.

Maybe we could do the same to build trust in the results provided by LLVM's DA.
We could adapt Polly to print distance and direction vectors and then diff DA 
against Polly.

https://github.com/llvm/llvm-project/pull/140182
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to