Ugh, I lost the builddir, so give me an hour or so to get it back...

Please provide more details, certainly my x86_64-linux and i686-linux bootstraps went fine.

Note the most important part of the command line is `-march=znver2 -O3` .

Created attachment 57064 [details]
sanitizer_stack_store.ii

Crashes with:
```
# /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/build/./gcc/xgcc -c -shared-libgcc -B/var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/build/./gcc -nostdinc++ ./.libs/sanitizer_stack_store.ii -std=gnu++14 -m32 -O3 -march=znver2
/var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/libsanitizer/sanitizer_common/sanitizer_stack_store.cpp: In member function ‘__sanitizer::uptr* __sanitizer::StackStore::BlockInfo::GetOrUnpack(__sanitizer::StackStore*)’:
/var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/libsanitizer/sanitizer_common/sanitizer_stack_store.cpp:253:7: error: definition in block 31 does not dominate use in block 30
  253 | uptr *StackStore::BlockInfo::GetOrUnpack(StackStore *store) {
      |       ^~~~~~~~~~
for SSA_NAME: vect_value_122.379_589 in statement:
vect_value_122.379_590 = PHI <vect_value_122.379_589(30)>
PHI argument
vect_value_122.379_589
for PHI node
vect_value_122.379_590 = PHI <vect_value_122.379_589(30)>
during GIMPLE pass: vect
/var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/libsanitizer/sanitizer_common/sanitizer_stack_store.cpp:253:7: internal compiler error: verify_ssa failed
0x5604d3777a76 verify_ssa(bool, bool) [clone .constprop.0]
        /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/gcc/tree-ssa.cc:1203
0x5604d4e9b845 execute_function_todo
        /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/gcc/passes.cc:2095
0x5604d4e9b845 do_per_function
        /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/gcc/passes.cc:1687
0x5604d4e9b845 execute_todo
        /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/gcc/passes.cc:2142
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <https://bugs.gentoo.org/> for instructions.
```

```
# /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/build/./gcc/xgcc -v
Using built-in specs.
COLLECT_GCC=/var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/build/./gcc/xgcc
Target: x86_64-pc-linux-gnu
Configured with: /var/tmp/portage/sys-devel/gcc-14.0.1.9999/work/gcc-14.0.1.9999/configure --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu --prefix=/usr --bindir=/usr/x86_64-pc-linux-gnu/gcc-bin/14 --includedir=/usr/lib/gcc/x86_64-pc-linux-gnu/14/include --datadir=/usr/share/gcc-data/x86_64-pc-linux-gnu/14 --mandir=/usr/share/gcc-data/x86_64-pc-linux-gnu/14/man --infodir=/usr/share/gcc-data/x86_64-pc-linux-gnu/14/info --with-gxx-include-dir=/usr/lib/gcc/x86_64-pc-linux-gnu/14/include/g++-v14 --disable-silent-rules --disable-dependency-tracking --with-python-dir=/share/gcc-data/x86_64-pc-linux-gnu/14/python --enable-languages=c,c++,fortran --enable-obsolete --enable-secureplt --disable-werror --with-system-zlib --enable-nls --without-included-gettext --disable-libunwind-exceptions --enable-checking=yes,extra,rtl --with-bugurl=https://bugs.gentoo.org/ --with-pkgversion='Gentoo Hardened 14.0.1 p, commit 9d69e54a3b402b0fad067464bd402e92c14504a9' --with-gcc-major-version-only --enable-libstdcxx-time --enable-lto --disable-libstdcxx-pch --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-multilib --with-multilib-list=m32,m64 --disable-fixed-point --enable-targets=all --enable-libgomp --disable-libssp --disable-libada --enable-cet --disable-systemtap --enable-valgrind-annotations --disable-vtable-verify --disable-libvtv --with-zstd --with-isl --disable-isl-version-check --enable-default-pie --enable-host-pie --enable-host-bind-now --enable-default-ssp --with-build-config='bootstrap-O3 bootstrap-lto bootstrap-cet'
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 14.0.1 20240112 (experimental) 8b447fa89d5e6b052c9309dfd1dafebdbd829ff9 (Gentoo Hardened 14.0.1 p, commit 9d69e54a3b402b0fad067464bd402e92c14504a9)
```

Still happens at r14-7220-gac1a399bf61b04 .

Let me try to reduce it.

Created attachment 57065 [details]
Reduced testcase

`-O3 -march=znver2 ` is enough with this testcase to reproduce it.

Note the testcase also ICEs on aarch64 with just -O3.

Hmm curious, does it work for you with --with-build-config='bootstrap-O3' that's how I tested it before

vect_create_epilog_for_reduction needs to handle the case where the vectorizer has picked a different exit than the main one.

diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
index eccf0953bba..6f761a4a78f 100644
--- a/gcc/tree-vect-loop.cc
+++ b/gcc/tree-vect-loop.cc
@@ -5965,7 +5965,8 @@ vect_create_epilog_for_reduction (loop_vec_info loop_vinfo,
      loop-closed PHI of the inner loop which we remember as
      def for the reduction PHI generation.  */
   bool double_reduc = false;
-  bool main_exit_p = LOOP_VINFO_IV_EXIT (loop_vinfo) == loop_exit;
+  bool main_exit_p = LOOP_VINFO_IV_EXIT (loop_vinfo) == loop_exit
+                    && !LOOP_VINFO_EARLY_BREAKS_VECT_PEELED (loop_vinfo);
   stmt_vec_info rdef_info = stmt_info;
   if (STMT_VINFO_DEF_TYPE (stmt_info) == vect_double_reduction_def)
     {

fixes it. But would be good if I can reproduce the bootstrap issue. Will try with provided options.

(In reply to Tamar Christina from comment #8)
> Hmm curious, does it work for you with --with-build-config='bootstrap-O3'
> that's how I tested it before

Will have a look. I'll try ignore the bootstrap failure and build userland stuff now so we can get some other cases if any.

Did you have -march=<some ISA with good enough vector instructions> too? I imagine bare '-O2' or '-O3' will be ok on amd64.

I built with the above configure args but also:
1. make -j32 -l32 'STAGE1_CFLAGS=-O3 -march=znver2 -pipe' 'BOOT_CFLAGS=-O3 -march=znver2 -pipe' ...
2. CFLAGS=... exported in the env

Like in PR113374, started to ICE with one message in                                                                                                                                  
r14-7194-g6cb155a6cf314232248a12bdd395ed4151ae5a28, another one in                                                                                                                    
r14-7195-g411de96dbf2bdafc7a90ebbfc63e68afd6388d29 and yet another one                                                                                                                
r14-7196-g99c0a540d6689ede068f9ba98af6f38c3cd71362 which is what we get until                                                                                                         
latest trunk.

Re-confirmed after my fixes btw.

Yes I had to rerun my baseline after updating trunk. Will post patch once peak finishes

Yes I had to rerun my baseline after updating trunk. Will post patch once peak finishes

Ok, the fix fixes the ICE but after rebasing to trunk I get a misscompile during bootstrap which miscompiles the x86 backend.

This is likely related to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113539 so tracking it down...

Ok, I've submitted the patch since the ICE and miscompare are unrelated.

I'll keep this ticket open in any case.  The miscompares didn't happen based on commits from ~2 weeks ago, So this will give me a place to start.

Hopefully send a patch for those tomorrow.

The master branch has been updated by Tamar Christina <tnfchris@gcc.gnu.org>:

https://gcc.gnu.org/g:72429448fd16733f876b282bb37a0653049c446d

commit r14-8382-g72429448fd16733f876b282bb37a0653049c446d
Author: Tamar Christina <tamar.christina@arm.com>
Date:   Wed Jan 24 07:37:17 2024 +0000

    middle-end: fix epilog reductions when vector iters peeled [PR113364]
    
    This fixes a bug where vect_create_epilog_for_reduction does not handle the
    case where all exits are early exits.  In this case we should do like induction
    handling code does and not have a main exit.
    
    This shows that some new miscompiles are happening (stage3 is likely miscompiled)
    but that's unrelated to this patch and I'll look at it next.
    
    gcc/ChangeLog:
    
            PR tree-optimization/113364
            * tree-vect-loop.cc (vect_create_epilog_for_reduction): If all exits all
            early exits then we must reduce from the first offset for all of them.
    
    gcc/testsuite/ChangeLog:
    
            PR tree-optimization/113364
            * gcc.dg/vect/vect-early-break_107-pr113364.c: New test.

*** Bug 113561 has been marked as a duplicate of this bug. ***

*** Bug 113555 has been marked as a duplicate of this bug. ***

I'll continue with the miscompile in PR113539 which I think is a duplicate of PR113467 which I'll verify now.