Re: [r17-126 Regression] FAIL: g++.dg/tree-ssa/cond-add-vec-2.C -std=gnu++98 scan-tree-dump-times forwprop1 "VEC_COND_EXPR" 2 on Linux/x86_64
Andrew Pinski via Gcc-regression <[email protected]>
| Newsgroups | gmane.comp.gcc.regression,gmane.comp.gcc.patches |
|---|---|
| Message-ID | <CALvbMcDYrRuitHeM5k7WV=2g0tefyu0L7nzm0-xKgCBYazdxjg@mail.gmail.com> |
On Tue, Apr 28, 2026 at 10:36 PM Haochen Jiang <[email protected]> wrote: > > On Linux/x86_64, > > 5e26bbfbdfa875ef464896bb8768669dd5a9ceb9 is the first bad commit > commit 5e26bbfbdfa875ef464896bb8768669dd5a9ceb9 > Author: Netanel Komm <[email protected]> > Date: Tue Apr 28 01:59:37 2026 +0300 > > match.pd: x != CST1 ? x + CST2 : CST3 -> x + CST2 [PR112659, PR122996] > > caused > > FAIL: g++.dg/tree-ssa/cond-add-vec-2.C -std=gnu++20 scan-tree-dump-times forwprop1 "VEC_COND_EXPR" 2 > FAIL: g++.dg/tree-ssa/cond-add-vec-2.C -std=gnu++26 scan-tree-dump-times forwprop1 "VEC_COND_EXPR" 2 > FAIL: g++.dg/tree-ssa/cond-add-vec-2.C -std=gnu++98 scan-tree-dump-times forwprop1 "VEC_COND_EXPR" 2 > > with GCC configured with > > ../../gcc/configure --prefix=/export/users3/haochenj/src/gcc-bisect/master/master/r17-126/usr --enable-clocale=gnu --with-system-zlib --with-demangler-in-ld --with-fpmath=sse --enable-languages=c,c++,fortran --enable-cet --without-isl --enable-libmpx x86_64-linux --disable-bootstrap > > To reproduce: > > $ cd {build_dir}/gcc && make check RUNTESTFLAGS="dg.exp=g++.dg/tree-ssa/cond-add-vec-2.C --target_board='unix{-m32\ -march=cascadelake}'" > $ cd {build_dir}/gcc && make check RUNTESTFLAGS="dg.exp=g++.dg/tree-ssa/cond-add-vec-2.C --target_board='unix{-m64\ -march=cascadelake}'" I will fix this, this is just a testsuite issue in this case. VEC_COND_EXPR does not show up but instead COND_ADD. There is another missed optimization due to COND_ADD which I will file separately too. Thanks, Andrew > > (Please directly reply to this email for question about this report.) > (If you met problems with cascadelake related, disabling AVX512F in command line might save that.) > (However, please make sure that there is no potential problems with AVX512.)