I built the LLVM and Clang (version 3.2) on Windows with the help of CMake and MinGW. The building is easy and successful. However, Clang failed to work with the sample code.
#include <stdarg.h>
#include <stdio.h>
int main()
{
printf("BAD: %lld\n", 1);
return 0;
}
When I compiled it with clang as
clang -o printf.exe printf.c -v
On Windows, it failed with messages
clang version 3.2 (branches/release_32 172788)
Target: i686-pc-mingw32
Thread model: posix
"D:/llvm/Build/bin/clang.exe" -cc1 -triple i686-pc-mingw32 -S -disable-free -main-file-name printf.c -mrelocation-model static -mdisable-fp-elim -fmath-errno -mconstructor-aliases -target-cpu pentium4 -momit-leaf-frame-pointer -v -resource-dir "D:/llvm/Build/bin\\..\\lib\\clang\\3.2" -fmodule-cache-path "C:\\Users\\usrname\\AppData\\Local\\Temp\\clang-module-cache" -fno-dwarf-directory-asm -ferror-limit 19 -fmessage-length 140 -mstackrealign -fno-use-cxa-atexit -fobjc-runtime=gcc -fdiagnostics-show-option -fcolor-diagnostics -o C:/Users/usrname/AppData/Local/Temp/printf-976141.s -x c printf.c
clang -cc1 version 3.2 based upon LLVM 3.2svn default target i686-pc-mingw32
ignoring nonexistent directory "/usr/local/include"
ignoring nonexistent directory "D:/llvm/Build/bin/../lib/clang/3.2/../../../i686-w64-mingw32/include"
ignoring nonexistent directory "D:/llvm/Build/bin/../lib/clang/3.2/../../../x86_64-w64-mingw32/include"
ignoring nonexistent directory "/mingw/include"
ignoring nonexistent directory "/usr/include"
#include "..." search starts here:
#include search starts here:
D:/llvm/Build/bin/../lib/clang/3.2/include
D:/llvm/Build/bin/../lib/clang/3.2/../../../include
c:/mingw/include
End of search list.
printf.c:6:24: warning: format specifies type 'long long' but the argument has type 'int' [-Wformat]
printf("BAD: %lld\n", 1);
~~~~ ^
%d
1 warning generated.
"C:/MinGW/bin/gcc.exe" -v -c -m32 -o C:/Users/usrname/AppData/Local/Temp/printf-976142.o -x assembler C:/Users/usrname/AppData/Local/Temp/printf-976141.s
Using built-in specs.
COLLECT_GCC=C:/MinGW/bin/gcc.exe
COLLECT_LTO_WRAPPER=c:/mingw/bin/../libexec/gcc/mingw32/4.6.1/lto-wrapper.exe
Target: mingw32
Configured with: ../gcc-4.6.1/configure --enable-languages=c,c++,fortran,objc,obj-c++ --disable-sjlj-exceptions --with-dwarf2 --enable-shared --enable-libgomp --disable-win32-registry --enable-libstdcxx-debug --enable-version-specific-runtime-libs --build=mingw32 --prefix=/mingw
Thread model: win32
gcc version 4.6.1 (GCC)
COLLECT_GCC_OPTIONS='-v' '-c' '-m32' '-o' 'C:/Users/usrname/AppData/Local/Temp/printf-976142.o' '-mtune=i386' '-march=i386'
c:/mingw/bin/../lib/gcc/mingw32/4.6.1/../../../../mingw32/bin/as.exe -o C:/Users/usrname/AppData/Local/Temp/printf-976142.o C:/Users/usrname/AppData/Local/Temp/printf-976141.s
C:/Users/usrname/AppData/Local/Temp/printf-976141.s: Assembler messages:
C:/Users/usrname/AppData/Local/Temp/printf-976141.s:7: Error: bad expression
C:/Users/usrname/AppData/Local/Temp/printf-976141.s:7: Error: junk at end of line, first unrecognized character is `\'
clang: error: assembler (via gcc) command failed with exit code 1 (use -v to see invocation)
The generated temporary assemble file is:
.def _main;
.scl 2;
.type 32;
.endef
.text
.globl _main
.align 16, 0x90, "\357\276\255\336"
_main:
It seems that Clang on Windows generats a wrong format assemble file.
But on Linux, it generats an obejct file directly instead of an assemble file and succeeds to compile.
How can I fix this problem? Thanks very much!
There is something wrong with your installation.
Please find my Clang builds here:
Clang package and required GCC package.
As shown in the Clang package download directory, extract both to the same directory, and add "mingw32-dw2/bin" to PATH. I tested you code and it works.
Note your code has an error:
#include <stdarg.h>
#include <stdio.h>
int main()
{
printf("BAD: %lld\n", 1LL); // NOTE the suffix specifying "long long"
return 0;
}
Observe the line of the failing invocation
"C:/MinGW/bin/gcc.exe" -v -c -m32 -o C:/Users/usrname/AppData/Local/Temp/printf-976142.o -x assembler C:/Users/usrname/AppData/Local/Temp/printf-976141.s
clang defaults to AT&T syntax and not Intel, if I read the docs correctly. For your version of gcc the arguments may require looking into for missing/incorrect ones. If running successful build has priority then alternatively try compiling to .o, then linking to be an exe with llvm's lld first, and then gnu's ld second.
Related
I have built from scratch Clang on MacOS and I am having problems with it.
Used following to configure Clang for building:
cmake -G "Unix Makefiles" -DCMAKE_INSTALL_PREFIX=/opt/clang-12 -DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD="AArch64;ARM;X86" -DLLVM_ENABLE_PROJECTS="clang;libcxx;libcxxabi;lldb" ../llvm
This is my test program:
#include <ostream>
#include <algorithm>
int main(int argc, char** argv) {
return 0;
}
OK if compiled with Clang 11.0.3 that ships with XCode clang++ -c test.cc
ERROR if using new Clang 12 (built) /opt/clang-12/bin/clang++ -c test.cc
In file included from test.cc:2:
In file included from /opt/clang-12/bin/../include/c++/v1/ostream:138:
In file included from /opt/clang-12/bin/../include/c++/v1/ios:214:
In file included from /opt/clang-12/bin/../include/c++/v1/iosfwd:95:
/opt/clang-12/bin/../include/c++/v1/wchar.h:119:15: fatal error: 'wchar.h' file not found
#include_next <wchar.h>
^~~~~~~~~
1 error generated.
Depending on what I am trying to include I get different include errors.
➜ tests /opt/clang-12/bin/clang++ -v -c test.cc
clang version 12.0.0 (https://github.com/llvm/llvm-project.git b529c5270c99e0ca18e3cbd9a5f50eb0970e560a)
Target: x86_64-apple-darwin19.6.0
Thread model: posix
InstalledDir: /opt/clang-12/bin
(in-process)
"/opt/clang-12/bin/clang-12" -cc1 -triple x86_64-apple-macosx10.15.0 -Wundef-prefix=TARGET_OS_ -Werror=undef-prefix -Wdeprecated-objc-isa-usage -Werror=deprecated-objc-isa-usage -emit-obj -mrelax-all --mrelax-relocations -disable-free -disable-llvm-verifier -discard-value-names -main-file-name test.cc -mrelocation-model pic -pic-level 2 -mframe-pointer=all -fno-rounding-math -munwind-tables -fcompatibility-qualified-id-block-type-checking -target-cpu penryn -debugger-tuning=lldb -target-linker-version 556.6 -v -resource-dir /opt/clang-12/lib/clang/12.0.0 -stdlib=libc++ -internal-isystem /opt/clang-12/bin/../include/c++/v1 -internal-isystem /usr/include/c++/v1 -internal-isystem /usr/local/include -internal-isystem /opt/clang-12/lib/clang/12.0.0/include -internal-externc-isystem /usr/include -fdeprecated-macro -fdebug-compilation-dir /Users/duddie/Projects/Audio/1voct/1voct_modular/tests -ferror-limit 19 -stack-protector 1 -fblocks -fencode-extended-block-signature -fregister-global-dtors-with-atexit -fgnuc-version=4.2.1 -fcxx-exceptions -fexceptions -fmax-type-align=16 -fcolor-diagnostics -o test.o -x c++ test.cc
clang -cc1 version 12.0.0 based upon LLVM 12.0.0git default target x86_64-apple-darwin19.6.0
ignoring nonexistent directory "/usr/include/c++/v1"
ignoring nonexistent directory "/usr/include"
#include "..." search starts here:
#include <...> search starts here:
/opt/clang-12/bin/../include/c++/v1
/usr/local/include
/opt/clang-12/lib/clang/12.0.0/include
/System/Library/Frameworks (framework directory)
/Library/Frameworks (framework directory)
End of search list.
In file included from test.cc:2:
In file included from /opt/clang-12/bin/../include/c++/v1/ostream:138:
In file included from /opt/clang-12/bin/../include/c++/v1/ios:214:
In file included from /opt/clang-12/bin/../include/c++/v1/iosfwd:95:
/opt/clang-12/bin/../include/c++/v1/wchar.h:119:15: fatal error: 'wchar.h' file not found
#include_next <wchar.h>
^~~~~~~~~
1 error generated.
ERROR if Clang 11.0.3 (XCode) or Clang 12 but target for ARM clang++ -target arm-none-eabi -c test.cc
test.cc:2:10: fatal error: 'ostream' file not found
#include <ostream>
^~~~~~~~~
1 error generated.
Any ideas what am I doing wrong?
Add the following to the shell profile file ~/.bash_profile or ~/.zsh_profile or any other way to set environment variable accessible to the shell you're running it in.
export SDKROOT="/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk"
Verify the correct path to the SDK on your macOS/Xcode version.
Reopen Terminal (sourcing the file may not work reliably).
Then try your compile command again.
/opt/clang-12/bin/clang++ -c test.cc
The Command Line Tools package installs the macOS system headers
inside the macOS SDK. Software that compiles with the installed tools
will search for headers within the macOS SDK provided by either Xcode
at:
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk
or the Command Line Tools at:
/Library/Developer/CommandLineTools/SDKs/MacOSX.sdk depending on which
is selected using xcode-select. The command line tools will search the
SDK for system headers by default. However, some software may fail to
build correctly against the SDK and require macOS headers to be
installed in the base system under /usr/include. If you are the
maintainer of such software, we encourage you to update your project
to work with the SDK or file a bug report for issues that are
preventing you from doing so. As a workaround, an extra package is
provided which will install the headers to the base system. In a
future release, this package will no longer be provided. You can find
this package at:
/Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg
https://developer.apple.com/documentation/xcode-release-notes/xcode-10-release-notes
Probably a more portable one (I found the hint at llvm-project/llvm/utils/lit/lit/util.py:399):
export SDKROOT=$(xcrun --show-sdk-path --sdk macosx)
Also, you can also add -isysroot argument to clang driver to alter the default system root for header searching:
clang++ a.cpp -isysroot $(xcrun --show-sdk-path --sdk macosx)
More details and investigation notes here
For what it's worth, I was able to fix the problem with MacOs Big Sur by doing
export SDKROOT='/Library/Developer/CommandLineTools/SDKs/MacOSX11.1.sdk'
I've been trying to install the OpenMP 4.5 off-loading to Nvidia GPU version of gcc for a while and so far no success, although I'm getting closer.
This time, I followed this script, where I have made two changes: First I specified the trunk version of gcc instead of 7.2, secondly nvptx-newlib is now included in nvptx-tools according to the github repository, so I removed that part of the script. For easy reference, the original script is
#!/bin/sh
#
# Build GCC with support for offloading to NVIDIA GPUs.
#
work_dir=$HOME/offload/wrk
install_dir=$HOME/offload/install
# Location of the installed CUDA toolkit
cuda=/usr/local/cuda
# Build assembler and linking tools
mkdir -p $work_dir
cd $work_dir
git clone https://github.com/MentorEmbedded/nvptx-tools
cd nvptx-tools
./configure \
--with-cuda-driver-include=$cuda/include \
--with-cuda-driver-lib=$cuda/lib64 \
--prefix=$install_dir
make
make install
cd ..
# Set up the GCC source tree
git clone https://github.com/MentorEmbedded/nvptx-newlib
svn co svn://gcc.gnu.org/svn/gcc/tags/gcc_7_2_0_release gcc
cd gcc
contrib/download_prerequisites
ln -s ../nvptx-newlib/newlib newlib
cd ..
target=$(gcc/config.guess)
# Build nvptx GCC
mkdir build-nvptx-gcc
cd build-nvptx-gcc
../gcc/configure \
--target=nvptx-none --with-build-time-tools=$install_dir/nvptx-none/bin \
--enable-as-accelerator-for=$target \
--disable-sjlj-exceptions \
--enable-newlib-io-long-long \
--enable-languages="c,c++,fortran,lto" \
--prefix=$install_dir
make -j4
make install
cd ..
# Build host GCC
mkdir build-host-gcc
cd build-host-gcc
../gcc/configure \
--enable-offload-targets=nvptx-none \
--with-cuda-driver-include=$cuda/include \
--with-cuda-driver-lib=$cuda/lib64 \
--disable-bootstrap \
--disable-multilib \
--enable-languages="c,c++,fortran,lto" \
--prefix=$install_dir
make -j4
make install
cd ..
After quite a while, this successfully exits. Per the instructions on that webpage, I added $install_dir/lib64 to my LD_LIBRARY_PATH and additionally to LIBRARY_PATH.
Then as a test, I have the following basic test program
#include <omp.h>
#include <cmath>
#include <iostream>
int main()
{
double data_array[1000000];
#pragma omp target teams distribute
for (int idx = 0; idx < 1000000; ++idx)
{
data_array[idx] = idx;
}
std::cout << "Hopefully this ran on the gpu...\n";
}
Then I try to compile this using offload/install/bin/g++ -fopenmp -foffload=nvptx-none main.cpp then it returns with the following error message:
x86_64-pc-linux-gnu-accel-nvptx-none-gcc: error: libgomp.spec: No such file or directory
mkoffload: fatal error: offload/install/bin/x86_64-pc-linux-gnu-accel-nvptx-none-gcc returned 1 exit status
compilation terminated.
lto-wrapper: fatal error: /home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0//accel/nvptx-none/mkoffload returned 1 exit status
compilation terminated.
/usr/bin/ld: error: lto-wrapper failed
collect2: error: ld returned 1 exit status
The file libgomp.spec can be found in the aforementioned $install_dir/lib64, which on my system is offload/install/lib64/.
Some more information about my system:
Ubuntu 16.04, accessed through slurm
Cuda 9.0.176
4x Nvidia Tesla V100
offload/install/bin/g++ -v reports:
Using built-in specs.
COLLECT_GCC=offload/install/bin/g++
COLLECT_LTO_WRAPPER=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/lto-wrapper
OFFLOAD_TARGET_NAMES=nvptx-none
Target: x86_64-pc-linux-gnu
Configured with: ../gcc/configure --enable-offload-targets=nvptx-none --with-cuda-driver-include=/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/include --with-cuda-driver-lib=/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64 --disable-bootstrap --disable-multilib --enable-languages=c,c++,fortran,lto --prefix=/home/over_ng/offload/install
Thread model: posix
gcc version 9.0.0 20180627 (experimental) (GCC)
offload/install/bin/g++ -print-search-dirs reports
install: /home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/
programs: =/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/bin/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/bin/x86_64-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/bin/
libraries: =/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/x86_64-pc-linux-gnu/9.0.0/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/x86_64-linux-gnu/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/../lib64/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib/x86_64-pc-linux-gnu/9.0.0/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib/x86_64-linux-gnu/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib/../lib64/:/home/over_ng/offload/install/lib64/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib64/x86_64-linux-gnu/:/home/over_ng/offload/install/lib64/../lib64/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/lib/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/lib/x86_64-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/lib/../lib64/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../x86_64-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../lib64/:/lib/x86_64-pc-linux-gnu/9.0.0/:/lib/x86_64-linux-gnu/:/lib/../lib64/:/usr/lib/x86_64-pc-linux-gnu/9.0.0/:/usr/lib/x86_64-linux-gnu/:/usr/lib/../lib64/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib/:/home/over_ng/offload/install/lib64/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/lib/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../:/lib/:/usr/lib/
And finally, offload/install/bin/g++ -fopenmp -foffload=nvptx-none -v main.cpp reports
Using built-in specs.
COLLECT_GCC=offload/install/bin/g++
COLLECT_LTO_WRAPPER=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/lto-wrapper
OFFLOAD_TARGET_NAMES=nvptx-none
Target: x86_64-pc-linux-gnu
Configured with: ../gcc/configure --enable-offload-targets=nvptx-none --with-cuda-driver-include=/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/include --with-cuda-driver-lib=/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64 --disable-bootstrap --disable-multilib --enable-languages=c,c++,fortran,lto --prefix=/home/over_ng/offload/install
Thread model: posix
gcc version 9.0.0 20180627 (experimental) (GCC)
COLLECT_GCC_OPTIONS='-fopenmp' '-foffload=nvptx-none' '-v' '-shared-libgcc' '-mtune=generic' '-march=x86-64' '-pthread'
/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/cc1plus -quiet -v -imultiarch x86_64-linux-gnu -D_GNU_SOURCE -D_REENTRANT main.cpp -quiet -dumpbase main.cpp -mtune=generic -march=x86-64 -auxbase main -version -fopenmp -foffload=nvptx-none -o /tmp/cc9FAd0p.s
GNU C++14 (GCC) version 9.0.0 20180627 (experimental) (x86_64-pc-linux-gnu)
compiled by GNU C version 8.1.0, GMP version 6.1.0, MPFR version 3.1.4, MPC version 1.0.3, isl version isl-0.18-GMP
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
ignoring nonexistent directory "/usr/local/include/x86_64-linux-gnu"
ignoring nonexistent directory "/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../x86_64-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/include
/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/include
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../include/c++/9.0.0
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../include/c++/9.0.0/x86_64-pc-linux-gnu
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../include/c++/9.0.0/backward
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/include
/usr/local/include
/home/over_ng/offload/install/include
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/include-fixed
/usr/include/x86_64-linux-gnu
/usr/include
End of search list.
GNU C++14 (GCC) version 9.0.0 20180627 (experimental) (x86_64-pc-linux-gnu)
compiled by GNU C version 8.1.0, GMP version 6.1.0, MPFR version 3.1.4, MPC version 1.0.3, isl version isl-0.18-GMP
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
Compiler executable checksum: 716ed3567afb9cd0b736d2b474553211
COLLECT_GCC_OPTIONS='-fopenmp' '-foffload=nvptx-none' '-v' '-shared-libgcc' '-mtune=generic' '-march=x86-64' '-pthread'
as -v --64 -o /tmp/cc2TYtU2.o /tmp/cc9FAd0p.s
GNU assembler version 2.26.1 (x86_64-linux-gnu) using BFD version (GNU Binutils for Ubuntu) 2.26.1
COMPILER_PATH=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/
LIBRARY_PATH=/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/../lib64/:/home/over_ng/offload/install/lib64/../lib64/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../lib64/:/lib/x86_64-linux-gnu/:/lib/../lib64/:/usr/lib/x86_64-linux-gnu/:/usr/lib/../lib64/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/:/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib/:/home/over_ng/offload/install/lib64/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../:/lib/:/usr/lib/
Reading specs from /home/over_ng/offload/install/lib64/../lib64/libgomp.spec
COLLECT_GCC_OPTIONS='-fopenmp' '-foffload=nvptx-none' '-v' '-shared-libgcc' '-mtune=generic' '-march=x86-64' '-pthread'
/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/collect2 -plugin /home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/liblto_plugin.so -plugin-opt=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/lto-wrapper -plugin-opt=-fresolution=/tmp/ccnGrpRF.res -plugin-opt=-pass-through=-lgcc_s -plugin-opt=-pass-through=-lgcc -plugin-opt=-pass-through=-lpthread -plugin-opt=-pass-through=-lc -plugin-opt=-pass-through=-lgcc_s -plugin-opt=-pass-through=-lgcc --eh-frame-hdr -m elf_x86_64 -dynamic-linker /lib64/ld-linux-x86-64.so.2 /usr/lib/x86_64-linux-gnu/crt1.o /usr/lib/x86_64-linux-gnu/crti.o /home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/crtbegin.o /home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/crtoffloadbegin.o -L/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64/../lib64 -L/home/over_ng/offload/install/lib64/../lib64 -L/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0 -L/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../../../lib64 -L/lib/x86_64-linux-gnu -L/lib/../lib64 -L/usr/lib/x86_64-linux-gnu -L/usr/lib/../lib64 -L/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/cuda-9.0.176-m4ivnigh5kuty6u7tcnroxr5on5lot6s/lib64 -L/tools/spack/install/linux-ubuntu16.04-x86_64/gcc-5.4.0/subversion-1.9.7-f5fbcx4xhwzrq5rhhco7byj7cbx2f4fs/lib -L/home/over_ng/offload/install/lib64 -L/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/../../.. /tmp/cc2TYtU2.o -lstdc++ -lm -lgomp -lgcc_s -lgcc -lpthread -lc -lgcc_s -lgcc /home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/crtend.o /usr/lib/x86_64-linux-gnu/crtn.o /home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/crtoffloadend.o
/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/lto-wrapper -fresolution=/tmp/ccnGrpRF.res -flinker-output=exec -foffload-objects=/tmp/ccQDi0zV.ofldlist
/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0//accel/nvptx-none/mkoffload #/tmp/ccJAbpMz
offload/install/bin/x86_64-pc-linux-gnu-accel-nvptx-none-gcc #/tmp/ccoh8KPc
Using built-in specs.
COLLECT_GCC=offload/install/bin/x86_64-pc-linux-gnu-accel-nvptx-none-gcc
COLLECT_LTO_WRAPPER=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/lto-wrapper
Target: nvptx-none
Configured with: ../gcc/configure --target=nvptx-none --with-build-time-tools=/home/over_ng/offload/install/nvptx-none/bin --enable-as-accelerator-for=x86_64-pc-linux-gnu --disable-sjlj-exceptions --enable-newlib-io-long-long --enable-languages=c,c++,fortran,lto --prefix=/home/over_ng/offload/install
Thread model: single
gcc version 9.0.0 20180627 (experimental) (GCC)
COLLECT_GCC_OPTIONS='-v' '-m64' '-mgomp' '-v' '-fno-openacc' '-foffload-abi=lp64' '-fopenmp' '-o' '/tmp/ccNVxXFz.mkoffload'
/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/lto1 -quiet -dumpbase cc2TYtU2.o -m64 -mgomp -auxbase cc2TYtU2 -version -fno-openacc -foffload-abi=lp64 -fopenmp #/tmp/cchKIS8V -o /tmp/ccZLBhjz.s
GNU GIMPLE (GCC) version 9.0.0 20180627 (experimental) (nvptx-none)
compiled by GNU C version 8.1.0, GMP version 6.1.0, MPFR version 3.1.4, MPC version 1.0.3, isl version isl-0.18-GMP
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU GIMPLE (GCC) version 9.0.0 20180627 (experimental) (nvptx-none)
compiled by GNU C version 8.1.0, GMP version 6.1.0, MPFR version 3.1.4, MPC version 1.0.3, isl version isl-0.18-GMP
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
COLLECT_GCC_OPTIONS='-v' '-m64' '-mgomp' '-v' '-fno-openacc' '-foffload-abi=lp64' '-fopenmp' '-o' '/tmp/ccNVxXFz.mkoffload'
/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/../../../../../../nvptx-none/bin/as -o /tmp/ccRJFdvc.o /tmp/ccZLBhjz.s
COMPILER_PATH=/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/:/home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/:/home/over_ng/offload/install/libexec/gcc/nvptx-none/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/:/home/over_ng/offload/install/lib/gcc/nvptx-none/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/../../../../../../nvptx-none/bin/
LIBRARY_PATH=/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/mgomp/:/home/over_ng/offload/install/lib/gcc/x86_64-pc-linux-gnu/9.0.0/accel/nvptx-none/
Reading specs from libgomp.spec
x86_64-pc-linux-gnu-accel-nvptx-none-gcc: error: libgomp.spec: No such file or directory
mkoffload: fatal error: offload/install/bin/x86_64-pc-linux-gnu-accel-nvptx-none-gcc returned 1 exit status
compilation terminated.
lto-wrapper: fatal error: /home/over_ng/offload/install/libexec/gcc/x86_64-pc-linux-gnu/9.0.0//accel/nvptx-none/mkoffload returned 1 exit status
compilation terminated.
/usr/bin/ld: error: lto-wrapper failed
collect2: error: ld returned 1 exit status
On the same webpage as where I found the script, somebody else reported the same problem and reverting to gcc 7.2 is apparently a solution. Since I want to include the off-loading compiler in the Spack collection, I would like to be able to use any supported version. Although I can live with gcc 8 for the time being, as 9/trunk is still experimental.
This may imply that it is a bug in gcc, in that case I would like to report it to them!
Edit 1: As requested, a 'sane' CPU only program that seems to work fine:
#include <omp.h>
#include <cmath>
#include <vector>
#include <iostream>
int main()
{
const int size = 1000;
std::vector<double> sinTable(size);
#pragma omp parallel for
for(int n=0; n<size; ++n)
{
sinTable[n] = std::sin(2 * M_PI * n / size);
std::cout << sinTable[n] << '\n';
}
// the table is now initialized
}
This was compiled with offload/install/bin/g++ -fopenmp -v main_cpu.cpp -o cpu
I have been using the package gcc-offload-nvptx in the Ubuntu repository since Ubuntu 17.10. If I compile your test code like this
g++ test.cpp -fopenmp
I get a lto-wrapper failed error. This can be fixed using -fno-stack-protector like this
g++ test.cpp -fopenmp -fno-stack-protector
Then test code compiles and runs. You can see that it runs on the GPU using nvprof like this
sudo nvprof ./a.out
Some additional comments. In your test code I would use
#pragma omp target teams distribute parallel for
See OpenMP offloading to Nvidia wrong reduction
Also in your test code you should do something with data_array or the compiler might optimize your code away.
Ubuntu 18.04 also requires -fno-stack-protector.
I'm trying to install (py)caffe on ubuntu 17.10
However when I do make all I get the following error:
./include/caffe/common.hpp(84): error: namespace "std" has no member "isnan"
./include/caffe/common.hpp(85): error: namespace "std" has no member "isinf"
2 errors detected in the compilation of "/tmp/tmpxft_00004921_00000000-19_nesterov_solver.compute_61.cpp1.ii".
Makefile:594: recipe for target '.build_release/cuda/src/caffe/solvers/nesterov_solver.o' failed
or when I use cmake instead
/home/thijser/caffe/include/caffe/common.hpp(84): error: namespace "std" has no member "isnan"
/home/thijser/caffe/include/caffe/common.hpp(85): error: namespace "std" has no member "isinf"
2 errors detected in the compilation of "/tmp/tmpxft_00004e32_00000000-7_math_functions.cpp1.ii".
CMake Error at cuda_compile_1_generated_math_functions.cu.o.Release.cmake:282 (message):
Error generating file
/home/thijser/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/./cuda_compile_1_generated_math_functions.cu.o
Note that I'm not just going for sudo apt-get install caffe as doing so does not also install pycaffe, any solution that also installs pycaffe is also valid. I understand that isnan in std likely comes via boost or that this somehow involves a discrepancy between c++ versions? I'm not much of a c++ expert so I'm not sure how that works, I use gcc4.9 for this as nvcc doesn't like later versions of gcc. I know that caffe works on other people's machines and on a older version of ubuntu I had installed so that suggests it's all a configuration issue.
Anybody know how to fix this?
edit using verbose mode on cmake I was able to find out that it doing the following as it happened
/usr/bin/cmake -E remove /home/thijser/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/cuda_compile_1_generated_math_functions.cu.o.depend.tmp /home/thijser/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/cuda_compile_1_generated_math_functions.cu.o.NVCC-depend
-- Generating /home/thijser/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/./cuda_compile_1_generated_math_functions.cu.o
/usr/bin/nvcc /home/thijser/caffe/src/caffe/util/math_functions.cu -c -o /home/thijser/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/./cuda_compile_1_generated_math_functions.cu.o -ccbin /usr/bin/cc -m64 -DCAFFE_VERSION=1.0.0 -Xcompiler ,\"-fPIC\",\"-Wall\",\"-Wno-sign-compare\",\"-Wno-uninitialized\",\"-O3\",\"-DNDEBUG\" -gencode arch=compute_61,code=sm_61 -Xcudafe --diag_suppress=cc_clobber_ignored -Xcudafe --diag_suppress=integer_sign_change -Xcudafe --diag_suppress=useless_using_declaration -Xcudafe --diag_suppress=set_but_not_used -DUSE_LMDB -DUSE_LEVELDB -DUSE_OPENCV -DWITH_PYTHON_LAYER -Xcompiler -fPIC -DNVCC -I/home/thijser/caffe/include -I/home/thijser/caffe/src -I/usr/include -I/usr/include/hdf5/serial -I/usr/include/opencv -I/usr/include/x86_64-linux-gnu -I/usr/include/python2.7 -I/usr/local/lib/python2.7/dist-packages/numpy/core/include -I/home/thijser/caffe/build/include -I/home/thijser/caffe/build
/home/thijser/caffe/include/caffe/common.hpp(84): error: namespace "std" has no member "isnan"
/home/thijser/caffe/include/caffe/common.hpp(85): error: namespace "std" has no member "isinf"
testing against the following code:
https://ideone.com/Yxvt5m
$ gcc -std=c++11 test.cpp which is the same as g++ test.cpp
gave me
test.cpp: In function ‘int main()’:
test.cpp:7:15: error: ‘__builtin_isnan’ is not a member of ‘std’
cout << std::isnan(42.0) << std::isinf(42.0);
^
test.cpp:7:15: note: suggested alternative:
<built-in>: note: ‘__builtin_isnan’
test.cpp:7:35: error: ‘__builtin_isinf_sign’ is not a member of ‘std’
cout << std::isnan(42.0) << std::isinf(42.0);
^
test.cpp:7:35: note: suggested alternative:
<built-in>: note: ‘__builtin_isinf_sign’
suggesting an configuration issue on my machine, however clang++ test.cpp did not give any error suggesting that maybe we could use that to install caffe?
$ gcc --version
gcc (Ubuntu 4.8.5-4ubuntu6) 4.8.5
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.
$gcc -xc++ -E -v -
Using built-in specs.
COLLECT_GCC=gcc
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.8.5-4ubuntu6' --with-bugurl=file:///usr/share/doc/gcc-4.8/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.8 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.8 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-libmudflap --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.8-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --disable-werror --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.8.5 (Ubuntu 4.8.5-4ubuntu6)
COLLECT_GCC_OPTIONS='-E' '-v' '-mtune=generic' '-march=x86-64'
/usr/lib/gcc/x86_64-linux-gnu/4.8/cc1plus -E -quiet -v -imultiarch x86_64-linux-gnu -D_GNU_SOURCE - -mtune=generic -march=x86-64 -fstack-protector -Wformat -Wformat-security
ignoring duplicate directory "/usr/include/x86_64-linux-gnu/c++/4.8"
ignoring nonexistent directory "/usr/local/include/x86_64-linux-gnu"
ignoring nonexistent directory "/usr/lib/gcc/x86_64-linux-gnu/4.8/../../../../x86_64-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
/usr/include/c++/4.8
/usr/include/x86_64-linux-gnu/c++/4.8
/usr/include/c++/4.8/backward
/usr/lib/gcc/x86_64-linux-gnu/4.8/include
/usr/local/include
/usr/lib/gcc/x86_64-linux-gnu/4.8/include-fixed
/usr/include/x86_64-linux-gnu
/usr/include
End of search list.
Edit: thanks to #EliahKagan 's suggestions in the askubuntu chat I found that adding in the line CXXFLAGS += g++ -D_GLIBCXX_USE_C99_MATH -std=c++11 to the makefile solves this issue, however it inimitably throws the next error of
thijser#AI-core-Regain:~/caffe/build$ make all
[ 0%] Building CXX object src/caffe/CMakeFiles/caffeproto.dir/__/__/include/caffe/proto/caffe.pb.cc.o
In file included from /home/thijser/caffe/build/include/caffe/proto/caffe.pb.cc:5:
/home/thijser/caffe/build/include/caffe/proto/caffe.pb.h:7:10: fatal error:
'string' file not found
#include <string>
^~~~~~~~
1 error generated.
In my case, I also met this error when I built caffe on ubuntu 17.10, but I finally found solution:
Let add the following to:
/usr/include/x86_64-linux-gnu/c++/5/bits/c++config.h
/* #undef _GLIBCXX_USE_C99_MATH */
#define _GLIBCXX_USE_C99_MATH 1
After a long and detailed discussion with several of the people on the chat.askubuntu.com room #EliahKagan found out that there is a python3-caffe-cuda package which allows on to import caffe in python3. Note that this won't solve this issue for anyone in the future who has python2 but for me it works. In other words sudo apt-get install caffe and then using python3 rather then python2.7 worked for me.
Also note that we found that the underlying issue appears to be that gcc/g++ isn't finding the standard libraries (std) however it seems that clang and later version of gcc (gcc7)can find these libraries however these do not work with the current version of nvcc.
I am pretty sure you should set the right C++ standard. std::isnan is part of C++11 and newer GCCs use this version by default. Older ones still default to C++98.
In my file "six.cpp", I try to print a float, but the compiler keeps throwing an error. Here is the file "six.cpp":
#include <iostream>
#include <iomanip>
using namespace std;
int main(){
double f = 5.3;
cout << f << endl;
return 0;
}
Then when I try to compile using g++ six.cpp I get the following output using the -v tag.
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/lto-wrapper.exe
Target: x86_64-pc-cygwin
Configured with: /cygdrive/i/szsz/tmpp/gcc/gcc-4.9.3-1.x86_64/src/gcc-4.9.3/configure --srcdir=/cygdrive/i/szsz/tmpp/gcc/gcc-4.9.3-1.x86_64/src/gcc-4.9.3 --prefix=/usr --exec-prefix=/usr --localstatedir=/var --sysconfdir=/etc --docdir=/usr/share/doc/gcc --htmldir=/usr/share/doc/gcc/html -C --build=x86_64-pc-cygwin --host=x86_64-pc-cygwin --target=x86_64-pc-cygwin --without-libiconv-prefix --without-libintl-prefix --libexecdir=/usr/lib --enable-shared --enable-shared-libgcc --enable-static --enable-version-specific-runtime-libs --enable-bootstrap --enable-__cxa_atexit --with-dwarf2 --with-tune=generic --enable-languages=ada,c,c++,fortran,lto,objc,obj-c++ --enable-graphite --enable-threads=posix --enable-libatomic --enable-libgomp --disable-libitm --enable-libquadmath --enable-libquadmath-support --enable-libssp --enable-libada --enable-libgcj-sublibs --disable-java-awt --disable-symvers --with-ecj-jar=/usr/share/java/ecj.jar --with-gnu-ld --with-gnu-as --with-cloog-include=/usr/include/cloog-isl --without-libiconv-prefix --without-libintl-prefix --with-system-zlib --enable-linker-build-id
Thread model: posix
gcc version 4.9.3 (GCC)
COLLECT_GCC_OPTIONS='-v' '-mtune=generic' '-march=x86-64'
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/cc1plus.exe -quiet -v -Dunix -idirafter /usr/lib/gcc/x86_64-pc-cygwin/4.9.3/../../../../lib/../include/w32api -idirafter /usr/lib/gcc/x86_64-pc-cygwin/4.9.3/../../../../x86_64-pc-cygwin/lib/../lib/../../include/w32api six.cpp -quiet -dumpbase six.cpp -mtune=generic -march=x86-64 -auxbase six -version -o /tmp/ccfHbtlL.s
GNU C++ (GCC) version 4.9.3 (x86_64-pc-cygwin)
compiled by GNU C version 4.9.3, GMP version 6.0.0, MPFR version 3.1.2-p11, MPC version 1.0.3
warning: GMP header version 6.0.0 differs from library version 6.1.0.
warning: MPFR header version 3.1.2-p11 differs from library version 3.1.3.
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
ignoring nonexistent directory "/usr/local/include"
ignoring nonexistent directory "/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/../../../../x86_64-pc-cygwin/include"
ignoring duplicate directory "/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/../../../../x86_64-pc-cygwin/lib/../lib/../../include/w32api"
#include "..." search starts here:
#include <...> search starts here:
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/include/c++
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/include/c++/x86_64-pc-cygwin
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/include/c++/backward
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/include
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/include-fixed
/usr/include
/usr/lib/gcc/x86_64-pc-cygwin/4.9.3/../../../../lib/../include/w32api
End of search list.
GNU C++ (GCC) version 4.9.3 (x86_64-pc-cygwin)
compiled by GNU C version 4.9.3, GMP version 6.0.0, MPFR version 3.1.2-p11, MPC version 1.0.3
warning: GMP header version 6.0.0 differs from library version 6.1.0.
warning: MPFR header version 3.1.2-p11 differs from library version 3.1.3.
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
Compiler executable checksum: a7b6fac296390f5db29d753ab65194e7
six.cpp:6:3: internal compiler error: Illegal instruction
double f = 5.3;
^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
I just started programming in C++, and I'm very confused to as to why this is happening. I've also done a few google searches and have not found anything about this.
You've got an internal compiler error: Illegal instruction, which means the gcc compiler itself has failed. From the warnings it looks like you've mis-installed the gcc tool chain. Try removing it and reinstalling or upgrading.
The portion of source code where it failed is not especially interesting here, the important bit is "internal compiler error: Illegal instruction".
This means that your copy of g++ is built for a newer subarchitecture than the CPU you are running it on. For example, SSE4 instructions were introduced in 2006 and older CPU designs don't support them. Because SSE is used for floating-point calculations, it makes sense that finding a floating-point literal triggered the error, but the root cause is the instruction set mismatch.
If you check /proc/cpuinfo, you can find out what instruction set extensions are supported on your CPU, which will help you pick the right g++ build options (whether you build yourself or download a package built with those options).
Relevant options are the ones listed here, particularly -march=, -mfpmath=, and -m(no)fused-madd.
I am trying to build clang trunk on Ubuntu 16.04 and I am having build errors no matter what I try. First I built llvm/clang/libc++/libc++abi against gcc 5.4, this worked fine. Now I am trying to use the clang I just build to rebuild llvm/clang/libc++/libc++abi. This fails with the following error message:
[162/4396] Linking CXX executable bin/llvm-tblgen FAILED:
: && /usr/local/bin/clang++ -stdlib=libc++
-fPIC
-fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter
-Wwrite-strings -Wcast-qual -Wmissing-field-initializers
-pedantic -Wno-long-long -Wcovered-switch-default
-Wnon-virtual-dtor -Wdelete-non-virtual-dtor
-Werror=date-time -std=c++11 -fcolor-diagnostics
-ffunction-sections -fdata-sections -O3 -DNDEBUG
-lc++ -lc++abi -Wl,-allow-shlib-undefined
-Wl,-O3 -Wl,--gc-sections
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/AsmMatcherEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/AsmWriterEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/AsmWriterInst.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/Attributes.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CallingConvEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeEmitterGen.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenDAGPatterns.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenInstruction.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenMapTable.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenRegisters.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenSchedule.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CodeGenTarget.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DAGISelEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DAGISelMatcherEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DAGISelMatcherGen.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DAGISelMatcherOpt.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DAGISelMatcher.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DFAPacketizerEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DisassemblerEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/FastISelEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/FixedLenDecoderEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/InstrInfoEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/IntrinsicEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/OptParserEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/PseudoLoweringEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/RegisterInfoEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/SearchableTableEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/SubtargetEmitter.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/TableGen.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/X86DisassemblerTables.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/X86ModRMFilters.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/X86RecognizableInstr.cpp.o
utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/CTagsEmitter.cpp.o
-o bin/llvm-tblgen
lib/libLLVMSupport.a
lib/libLLVMTableGen.a
-lpthread
lib/libLLVMSupport.a
-lrt
-ldl
-ltinfo
-lpthread
-lz
-lm
lib/libLLVMDemangle.a
-Wl,-rpath,"\$ORIGIN/../lib" &&
: utils/TableGen/CMakeFiles/obj.llvm-tblgen.dir/DFAPacketizerEmitter.cpp.o: In function `(anonymous namespace)::DFAPacketizerEmitter::run(llvm::raw_ostream&)': /home/mehrlich/Code/llvm/llvm/utils/TableGen/DFAPacketizerEmitter.cpp:(.text._ZN12_GLOBAL__N_120DFAPacketizerEmitter3runERN4llvm11raw_ostreamE+0x253f): undefined reference to `std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>::~basic_string()' clang-4.0: error: linker command failed with exit code 1 (use -v to see invocation) [162/4396] Building CXX object lib/MC/CMakeFiles/LLVMMC.dir/MCContext.cpp.o ninja: build stopped: subcommand failed.
The relevant error being the undefined symbol:
In function (anonymous namespace)::DFAPacketizerEmitter::run(llvm::raw_ostream&)': /home/mehrlich/Code/llvm/llvm/utils/TableGen/DFAPacketizerEmitter.cpp:(.text._ZN12_GLOBAL__N_120DFAPacketizerEmitter3runERN4llvm11raw_ostreamE+0x253f): undefined reference to std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>::~basic_string()'
Presumably this is for one of two reasons:
1. Clang is using libstdc++ headers:
We can see from the failing command that clang++ picks up the -stdlib=libc++ switch correctly. Running % clang++ --stdlib=libc++ -E -x c++ - -v < /dev/null to inspect the default include paths with this options gives
clang version 4.0.0 (trunk 281213) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /usr/local/bin Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/5 Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/5.4.0 Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/6 Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/6.0.0 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/4.9 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/4.9.3 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5.4.0 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6.0.0 Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5.4.0 Candidate multilib: .;#m64 Selected multilib: .;#m64 Found CUDA installation: /usr/local/cuda, version 7.5 "/usr/local/bin/clang-4.0" -cc1 -triple x86_64-unknown-linux-gnu -E -disable-free -disable-llvm-verifier
-discard-value-names -main-file-name - -mrelocation-model static -mthread-model posix -mdisable-fp-elim -fmath-errno -masm-verbose -mconstructor-aliases -munwind-tables -fuse-init-array -target-cpu x86-64 -v -dwarf-column-info -debugger-tuning=gdb -resource-dir /usr/local/bin/../lib/clang/4.0.0 -internal-isystem /usr/local/bin/../include/c++/v1 -internal-isystem /usr/local/include
-internal-isystem /usr/local/bin/../lib/clang/4.0.0/include -internal-externc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include -internal-externc-isystem /usr/include -fdeprecated-macro -fdebug-compilation-dir /home/mehrlich/Code/llvm/build -ferror-limit 19 -fmessage-length 343
-fobjc-runtime=gcc -fcxx-exceptions -fexceptions -fdiagnostics-show-option -fcolor-diagnostics -o - -x c++ - clang -cc1 version 4.0.0 based upon LLVM 4.0.0svn default target x86_64-unknown-linux-gnu ignoring nonexistent directory "/include"
#include "..." search starts here:
#include <...> search starts here: /usr/local/bin/../include/c++/v1 /usr/local/include /usr/local/bin/../lib/clang/4.0.0/include /usr/include/x86_64-linux-gnu /usr/include End of search list.
# 1 "<stdin>"
# 1 "<built-in>" 1
# 1 "<built-in>" 3
# 328 "<built-in>" 3
# 1 "<command line>" 1
# 1 "<built-in>" 2
# 1 "<stdin>" 2
The first header search path #include <...> search starts here: /usr/local/bin/../include/c++/v1 is the correct location of libc++ headers. So I highly doubt it is this
2. Clang is not linking against libc++
Adding the options -lc++ -lc++abi to CMAKE_EXE_LINKER_FLAGS, CMAKE_SHARED_LINKER_FLAGS, CMAKE_MODULE_LINKER_FLAGS, and CMAKE_STATIC_LINKER_FLAGS gives this following error:
[65/4396] Linking CXX static library lib/libLLVMDemangle.a
FAILED: : && /usr/bin/cmake -E remove lib/libLLVMDemangle.a && /usr/bin/ar qc lib/libLLVMDemangle.a -lc++ -lc++abi lib/Demangle/CMakeFiles/LLVMDemangle.dir/ItaniumDemangle.cpp.o && /usr/bin/ranlib lib/libLLVMDemangle.a && :
/usr/bin/ar: invalid option -- '+'
Usage: /usr/bin/ar [emulation options] [-]{dmpqrstx}[abcDfilMNoPsSTuvV] [--plugin <name>] [member-name] [count] archive-file file...
/usr/bin/ar -M [<mri-script]
commands:
d - delete file(s) from the archive
m[ab] - move file(s) in the archive
p - print file(s) found in the archive
q[f] - quick append file(s) to the archive
r[ab][f][u] - replace existing or insert new file(s) into the archive
s - act as ranlib
t - display contents of archive
x[o] - extract file(s) from the archive
command specific modifiers:
[a] - put file(s) after [member-name]
[b] - put file(s) before [member-name] (same as [i])
[D] - use zero for timestamps and uids/gids (default)
[U] - use actual timestamps and uids/gids
[N] - use instance [count] of name
[f] - truncate inserted file names
[P] - use full path names when matching
[o] - preserve original dates
[u] - only replace files that are newer than current archive contents
generic modifiers:
[c] - do not warn if the library had to be created
[s] - create an archive index (cf. ranlib)
[S] - do not build a symbol table
[T] - make a thin archive
[v] - be verbose
[V] - display the version number
#<file> - read options from <file>
--target=BFDNAME - specify the target object format as BFDNAME
optional:
--plugin <p> - load the specified plugin
emulation options:
No emulation specific options
/usr/bin/ar: supported targets: elf64-x86-64 elf32-i386 elf32-iamcu elf32-x86-64 a.out-i386-linux pei-i386 pei-x86-64 elf64-l1om elf64-k1om elf64-little elf64-big elf32-little elf32-big pe-x86-64 pe-bigobj-x86-64 pe-i386 plugin srec symbolsrec verilog tekhex binary ihex
And we can see for this, that the linker options are being passed to /usr/bin/ar by cmake for static libraries (this seems like the wrong thing to do to me but I am not an expert in build processes so I might be wrong). Removing these options from CMAKE_STATIC_LINKER_FLAGS reverts to the original error message about undefined symbols, so presumably the llvm-tblgen target is a static library. llvm-tblgen is definitely an executable, not sure the CMAKE_EXE_LINKER_FLAGS are not being applied to it flags are being applied, which means I have no idea why this would not be working...
At this point I'm not sure who's fault this is, mine, clang, or cmake, but if anyone has any advice or expertise in doing this process, I'd appreciate the help.
Update
Now errors are in compiler-rt:
projects/compiler-rt/lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/sanitizer_unwind_linux_libcdep.cc.o: In function `__sanitizer::Unwind_GetIP(_Unwind_Context*)':
/home/mehrlich/Code/llvm/llvm/projects/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc:98: undefined reference to `_Unwind_GetIP'
projects/compiler-rt/lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/sanitizer_unwind_linux_libcdep.cc.o: In function `Unwind_GetIP':
/home/mehrlich/Code/llvm/llvm/projects/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc:98: undefined reference to `_Unwind_GetIP'
projects/compiler-rt/lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/sanitizer_unwind_linux_libcdep.cc.o: In function `__sanitizer::BufferedStackTrace::SlowUnwindStack(unsigned long, unsigned int)':
/home/mehrlich/Code/llvm/llvm/projects/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc:125: undefined reference to `_Unwind_Backtrace'
projects/compiler-rt/lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/sanitizer_unwind_linux_libcdep.cc.o: In function `SlowUnwindStack':
/home/mehrlich/Code/llvm/llvm/projects/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc:125: undefined reference to `_Unwind_Backtrace'
Seems related to this bug
This was caused by a recent bug in libc++ which accidentally removed the definition of ~basic_string() from versions of libc++.so built with GCC. The bug was introduced September 9th and I fixed it earlier today.
Unfortunately your libc++ installation is buggy and it will need to be updated before building Clang. You can simply install a new libc++ over top of the existing installation. This can probably be done using sudo make install-cxx from your current (buggy) build directory. Then you can continue building Clang as normal.
For more info about building/installing libc++ see http://libcxx.llvm.org/docs/BuildingLibcxx.html.
(Note: Fixing the libc++ installation should not require rebuilding LLVM or Clang.)