https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122971
Jeffrey A. Law <law at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|law at gcc dot gnu.org |unassigned at gcc dot
gnu.org
Last reconfirmed|2025-12-03 00:00:00 |2025-12-21
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
--- Comment #6 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Yea, I can trip this one, both on the commit referenced in the title as well as
on the tip of the trunk.
./xgcc -B./ -march=rv64gcv_zvl1024b -flto -mrvv-vector-bits=zvl
-mrvv-max-lmul=m2 -O3 -fsigned-char -fno-strict-aliasing -fwrapv j.c -c &&
riscv64-elf-gcc j.o && QEMU_CPU=rv64,vlen=1024,rvv_ta_all_1s=true
,rvv_ma_all_1s=true,v=true,zve32f=true,zve64f=true qemu-riscv64 ./a.out
That should give an output of "0" when the expected output is "16". Also note
-O3 triggers the bug while -O2 does not. -O2 doesn't have any vector code.
-O3 seems to vectorize just about everything.
