[Bug libstdc++/89927] New: Inconsistent behavior in std::regex when optimized
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89927 Bug ID: 89927 Summary: Inconsistent behavior in std::regex when optimized Product: gcc Version: 8.2.1 Status: UNCONFIRMED Severity: normal Priority: P3 Component: libstdc++ Assignee: unassigned at gcc dot gnu.org Reporter: xonar.leroux at gmail dot com Target Milestone: --- Created attachment 46073 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46073=edit Minimal case that triggers the bug Compiling with: `g++ test.cpp` and running results in no exception as expected. Compiling with `g++ -O3 test.cpp` and running results in terminate called after throwing an instance of 'std::regex_error' what(): Unexpected token in brace expression. Compiling with `clang++ -fsanitize=undefined test.cpp` and running /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/8.2.1/../../../../include/c++/8.2.1/bits/regex_compiler.h:200:40: runtime error: addition of unsigned offset to 0x5581d2265f61 overflowed to 0x5581d2264b15 terminate called after throwing an instance of 'std::regex_error' what(): Unexpected token in brace expression. I am using the latest g++, libstdc++11 and clang++ versions from Arch Linux (2019 April 2). GCC Version: Using built-in specs. COLLECT_GCC=g++ COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/lto-wrapper Target: x86_64-pc-linux-gnu Configured with: /build/gcc/src/gcc/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/ --enable-languages=c,c++,ada,fortran,go,lto,objc,obj-c++ --enable-shared --enable-threads=posix --enable-libmpx --with-system-zlib --with-isl --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --disable-libssp --enable-gnu-unique-object --enable-linker-build-id --enable-lto --enable-plugin --enable-install-libiberty --with-linker-hash-style=gnu --enable-gnu-indirect-function --enable-multilib --disable-werror --enable-checking=release --enable-default-pie --enable-default-ssp --enable-cet=auto Thread model: posix gcc version 8.2.1 20181127 (GCC)
[Bug c++/69289] Compiling without --profile-generate causes longer execution time (-O3)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69289 --- Comment #5 from Paul le roux --- On the GCC trunk version on godbolt.org there is still a difference between the two cases, but it seems to have turned around. The noprofile case has vector::resize partially inlined and in the profile case, it is not inlined. This makes more sense (to me at least) since I expect that certain functions won't be inlined so that one can gather better profile information. If it is the case that `--profile-generate` should be less aggressive when inlining then this bug report can be closed.
[Bug c++/81171] Segfault with auto template deduction
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81171 Paul le roux changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #2 from Paul le roux --- This has been fixed somewhere between when this was reported and 8.2.1 which I am currently using.
[Bug c++/81171] New: Segfault with auto template deduction
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81171 Bug ID: 81171 Summary: Segfault with auto template deduction Product: gcc Version: 7.1.1 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: xonar.leroux at gmail dot com Target Milestone: --- Created attachment 41608 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41608=edit Smallest case which still results in segfault Running with 'g++ gcc_bug.cpp' results in 'gcc_bug.cpp:8:1: internal compiler error: Segmentation fault' Replacing auto with char results in successful compilation. VERSIONS: g++ (GCC) 7.1.1 20170516 I also tested with GCC 5.4.0 and with -std=c++14 and it also segfaults.
[Bug c++/69289] Compiling without --profile-generate causes longer execution time (-O3)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69289 --- Comment #4 from Paul le roux --- I compiled gcc/g++ from svn and the slowdown is still present without the --profile-generate flag. (But both are faster than what they were :D )
[Bug c++/69289] New: Compiling without --profile-generate causes longer execution time (-O3)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69289 Bug ID: 69289 Summary: Compiling without --profile-generate causes longer execution time (-O3) Product: gcc Version: 5.3.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: xonar.leroux at gmail dot com Target Milestone: --- Created attachment 37349 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=37349=edit Source Running without --profile-generate (NOT --profile-use) causes a slowdown. The speed script that I run simply runs the program with time output. Effect is still present when using/not using -march=native and when using/not using -funroll-loops speed.sh 1 #!/bin/zsh 2 3 sudo cpupower frequency-set -g performance 4 5 time ./a.out 1 6 time ./a.out 2 7 time ./a.out 3 8 time ./a.out 4 9 time ./a.out 5 10 time ./a.out 1 11 time ./a.out 2 12 time ./a.out 3 13 time ./a.out 4 14 time ./a.out 5 15 16 sudo cpupower frequency-set -g powersave % g++ --version g++ (GCC) 5.3.0 Copyright (C) 2015 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. % cat /proc/version Linux version 4.3.3-2-ARCH (builduser@tobias) (gcc version 5.3.0 (GCC) ) #1 SMP PREEMPT Wed Dec 23 20:09:18 CET 2015 %g++ test.cpp -std=c++11 -O3 -funroll-loops -march=native --profile-generate -Wall -Wextra test.cpp:37:14: warning: unused parameter ‘argc’ [-Wunused-parameter] int main(int argc, char** argv) % ./speed.sh [sudo] password for XXX: Setting cpu: 0 Setting cpu: 1 Setting cpu: 2 Setting cpu: 3 Setting cpu: 4 Setting cpu: 5 Setting cpu: 6 Setting cpu: 7 166 ./a.out 1 0.32s user 0.55s system 99% cpu 0.875 total 166 ./a.out 2 0.33s user 0.54s system 100% cpu 0.870 total 166 ./a.out 3 0.33s user 0.54s system 100% cpu 0.870 total 166 ./a.out 4 0.33s user 0.55s system 99% cpu 0.874 total 166 ./a.out 5 0.35s user 0.52s system 99% cpu 0.870 total 166 ./a.out 1 0.32s user 0.55s system 99% cpu 0.867 total 166 ./a.out 2 0.35s user 0.52s system 99% cpu 0.872 total 166 ./a.out 3 0.32s user 0.56s system 99% cpu 0.889 total 166 ./a.out 4 0.33s user 0.53s system 99% cpu 0.868 total 166 ./a.out 5 0.32s user 0.55s system 100% cpu 0.873 total Setting cpu: 0 Setting cpu: 1 Setting cpu: 2 Setting cpu: 3 Setting cpu: 4 Setting cpu: 5 Setting cpu: 6 Setting cpu: 7 ./speed.sh 3.34s user 5.42s system 79% cpu 10.947 total % g++ test.cpp -std=c++11 -O3 -funroll-loops -march=native -Wall -Wextra test.cpp:37:14: warning: unused parameter ‘argc’ [-Wunused-parameter] int main(int argc, char** argv) % ./speed.sh Setting cpu: 0 Setting cpu: 1 Setting cpu: 2 Setting cpu: 3 Setting cpu: 4 Setting cpu: 5 Setting cpu: 6 Setting cpu: 7 166 ./a.out 1 0.43s user 0.56s system 100% cpu 0.990 total 166 ./a.out 2 0.42s user 0.57s system 99% cpu 0.983 total 166 ./a.out 3 0.41s user 0.58s system 99% cpu 0.997 total 166 ./a.out 4 0.47s user 0.54s system 99% cpu 1.005 total 166 ./a.out 5 0.43s user 0.56s system 99% cpu 0.988 total 166 ./a.out 1 0.45s user 0.54s system 99% cpu 0.991 total 166 ./a.out 2 0.42s user 0.55s system 99% cpu 0.971 total 166 ./a.out 3 0.39s user 0.59s system 99% cpu 0.982 total 166 ./a.out 4 0.46s user 0.52s system 99% cpu 0.985 total 166 ./a.out 5 0.48s user 0.51s system 99% cpu 0.985 total Setting cpu: 0 Setting cpu: 1 Setting cpu: 2 Setting cpu: 3 Setting cpu: 4 Setting cpu: 5 Setting cpu: 6 Setting cpu: 7 ./speed.sh 4.36s user 5.53s system 99% cpu 9.909 total
[Bug c++/69289] Compiling without --profile-generate causes longer execution time (-O3)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69289 Paul le roux changed: What|Removed |Added CC||xonar.leroux at gmail dot com --- Comment #1 from Paul le roux --- When copying the timings got switched around the --generate-profile is the faster one.
[Bug c++/69289] Compiling without --profile-generate causes longer execution time (-O3)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69289 --- Comment #2 from Paul le roux --- (In reply to Paul le roux from comment #1) > When copying the timings got switched around the --generate-profile is the > faster one. Ignore that. I just quickly looked at the total usage, but that includes the frequency-set. They are right way around.
