Hi, On Thu, May 10, 2018 at 10:57:16AM -0700, Carl Love wrote: > The following patch fixes issues found with the instruction counts for > the vsx-vector-6.h test. It was found that on a BE system where GCC is > configured with --with-cpu=power6 the expected counts in vsx-vector-6- > be.c do not match. This patch fixes the expected counts when > configuring and building GCC using --with-cpu=power6, --with-cpu=power7 > and --with-cpu=power8 on a BE system.
> 2018-05-10 Carl Love <c...@us.ibm.com> > * gcc.target/powerpc/vsx-vector-6-be.c (dg-options): Update to run on > Power 6 only. Then please rename it to vsx-vector-6-be.p6.c like the others. But, do we want a test for p6 at all? The test uses powerpc_vsx_ok which isn't true on a p6 (so it runs on p7 and up, but with -mcpu=power6; do we want to test that?) > (dg-final): Update xvcmpgtdp, xvcmpgedp counts for Power 6. > * gcc.target/powerpc/vsx-vector-6-be.p7.c (dg-final): New test file for > Power 7. > * gcc.target/powerpc/vsx-vector-6-be.p8.c (dg-final): New test file for > Power 8. Did you check that the updated counts make sense, i.e. that the expected code is generated? Okay for trunk if so (with the rename to .p6). Thanks! Segher
