https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117558
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2024-11-14
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
void
foo (double *restrict dest, double *src)
{
for (int i = 0; i < 2000; ++i)
dest[i] += src[i * 4];
}
is a testcase that shows the issue for both ARM SVE and RISC-V RVV, the former
vectorizing it only with fixed-length vectors when SLP is forced and the latter
not at all. W/o SLP (and thus without this check) we happily use VL vectors,
both targets using vector([2,2]) double.
