https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127587
--- Comment #2 from Andrew Stubbs <ams at gcc dot gnu.org> --- Claude says it's a bug in the testcase. It shows on GCN but not on x86 because x86 pages start clean and GCN pages don't. It suggests moving the initialization loop to before the first `tab' read.
