[Bug tree-optimization/111498] [14/15/16 regression] 951% profile quality regression between g:93996cfb308ffc63 (2023-09-18 03:40) and g:95d2ce05fb32e663 (2023-09-19 03:22)

[hubicka at ucw dot cz via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111498

--- Comment #6 from Jan Hubicka <hubicka at ucw dot cz> ---
> It creeps up slowly.  Whatever it exactly means.  Confirmed from the graph.
It tests that Kirhoff laws are met on basic block entry.  So sum of
counts of incoming edges should match count of BB.  The number reported
is sum of all disprepancies (so bigger differences counts more) with
profile feedback.