https://gcc.gnu.org/g:2563111c38bc59ba4a1fc60cbfd714d4bac94b15
commit r17-5653-g2563111c38bc59ba4a1fc60cbfd714d4bac94b15 Author: Igor Shevlyakov <[email protected]> Date: Tue Oct 27 14:32:59 2026 -0700 testsuite: Skip pr128080.c at -O0 The permutations in pr128080.c are merged only when optimizing, so at -O0 the test checks nothing. It is also slow there on some targets: on arm-eabi with -march=armv6s-m (Thumb-1) LRA spends more than ten minutes on the two 256-element shuffles that survive to RTL, and the compilation times out. Assisted-by: Claude Opus 5.5 gcc/testsuite/ChangeLog: * gcc.dg/torture/pr128080.c: Skip at -O0. Signed-off-by: Igor Shevlyakov <[email protected]> Diff: --- gcc/testsuite/gcc.dg/torture/pr128080.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/gcc/testsuite/gcc.dg/torture/pr128080.c b/gcc/testsuite/gcc.dg/torture/pr128080.c index cb0410c4997c..b84e6c1640ba 100644 --- a/gcc/testsuite/gcc.dg/torture/pr128080.c +++ b/gcc/testsuite/gcc.dg/torture/pr128080.c @@ -1,5 +1,7 @@ /* PR tree-optimization/128080 */ /* { dg-do run } */ +/* The permutations are merged only when optimizing. */ +/* { dg-skip-if "" { *-*-* } { "-O0" } { "" } } */ /* { dg-additional-options "-Wno-psabi" } */ typedef signed char V __attribute__((vector_size (256)));
