[PATCH] D72380: [DataFlow] Factor two worklist implementations out

Wed Jan 15 15:34:13 PST 2020

xazax.hun added a comment.

In D72380#1823019 <https://reviews.llvm.org/D72380#1823019>, @xazax.hun wrote:

> In D72380#1822927 <https://reviews.llvm.org/D72380#1822927>, @NoQ wrote:
>
> > The change in uninitialized values analysis gives me a bit of anxiety. Could you explain what exactly has changed that caused the change in the stats and why you think it doesn't make a difference, maybe give an example? (an example could be obtained by `creduce`-ing over "the stats have changed" criterion)
>
>
> We will process it multiple times with the original implementation, we traverse the CFG both using DFS and using reverse post order.

Sorry, I just realized that the original code initially considers all the nodes queued. In this case there is no extra pass. But it is still true that the original is mostly using DFS while the new one is only using RPO. So it comes down to the order we visit the basic blocks. Working on a minimal repro.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72380/new/

https://reviews.llvm.org/D72380