[llvm-dev] Extracting the names of the variables that creates loop-carried dependencies
David Blaikie via llvm-dev
llvm-dev at lists.llvm.org
Thu Aug 4 08:17:53 PDT 2016
Extracting things like variable names can be done in a best-effort basis by
examining debug info metadata/intrinsics in the LLVM IR - but, as I said,
it's totally best-effort (which is really "not much effort" when it comes
to optimized code) so it may not be present, accurate, etc.
On Wed, Aug 3, 2016 at 10:16 AM Miguel Aguilar via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
> I would like to know if it is possible to extract the source level names
> of variables that create loop-carried dependencies.
> For example, for the following code:
> for (int i = 0; i < A_ROW; i++)
> for (int j = 1; j < B_COL; j++)
> a_matrix[i][j] = a_matrix[i][j - 1];
> I get the following AST:
> #pragma omp parallel for
> for (int c0 = 0; c0 <= 31; c0 += 1)
> #pragma minimal dependence distance: 1
> for (int c1 = 0; c1 <= 30; c1 += 1)
> Stmt_for_body5(c0, c1);
> As expected the second loop has a dependence distance of 1. Then, here
> is when I would like to know if
> it possible to extract the name of the variable (a_matrix) e.g. by
> developing a custom LLVM pass, based on
> Polly passes.
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-dev