Created attachment 48845 [details] 32-bit sparc-sun-solaris2.11 vect-alias-check.c.163t.vect Between 20200707 (f75211822f8d84bb706421d3692e6c1cdbdfc3a8) and 20200708 (30fdaead5b7880c4e9f140618e26ad1c545642d5), two testcases regressed on Solaris/SPARC (both 32 and 64-bit): +FAIL: gcc.dg/vect/vect-alias-check.c -flto -ffat-lto-objects scan-tree-dump-times vect "vectorized 1 loops" 1 +FAIL: gcc.dg/vect/vect-alias-check.c scan-tree-dump-times vect "vectorized 1 loops" 1 +FAIL: gcc.dg/vect/vect-live-5.c -flto -ffat-lto-objects scan-tree-dump-times vect "vectorized 1 loops" 1 +FAIL: gcc.dg/vect/vect-live-5.c scan-tree-dump-times vect "vectorized 1 loops" 1 At the same time, an XPASS vanished: -XPASS: gcc.dg/vect/vect-alias-check-1.c -flto -ffat-lto-objects scan-tree-dump vect "using an address-based overlap test" -XPASS: gcc.dg/vect/vect-alias-check-1.c scan-tree-dump vect "using an address-based overlap test" There's also a report for the regressions on ARM. Dumps attached.
Created attachment 48846 [details] 32-bit sparc-sun-solaris2.11 vect-live-5.c.163t.vect
Quite some revs, two vectorizer changes. Do the FAILs still occur?
> --- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> --- > Quite some revs, two vectorizer changes. Do the FAILs still occur? Both still do.
As the regressed failures, it's highly suspected to be duplicated of PR96376.
Tested on Cfarm gcc102, confirmed it's due to 30fdaead5b7880c4e9f140618e26ad1c545642d5, marked as dup. *** This bug has been marked as a duplicate of bug 96376 ***