problems with gcc-versions when compiling c++ code.

rustyraptor · wrote on 6 Jun 2018, 18:20

I am not sure if this is a cuda issue or a qmake/qtcreator config issue but here is the output.

12:06:16: Running steps for project SuperRes...
12:06:16: Configuration unchanged, skipping qmake step.
12:06:16: Starting: "/usr/bin/make" 
Makefile:3001: warning: overriding recipe for target 'moc_mainviewer.cpp'
Makefile:2501: warning: ignoring old recipe for target 'moc_mainviewer.cpp'
Makefile:3221: warning: overriding recipe for target 'moc_mainwindow.cpp'
Makefile:2390: warning: ignoring old recipe for target 'moc_mainwindow.cpp'
Makefile:3679: warning: overriding recipe for target 'ui_mainwindow.h'
Makefile:3675: warning: ignoring old recipe for target 'ui_mainwindow.h'
Makefile:4638: warning: overriding recipe for target '../SuperRes/Obj/main.o'
Makefile:3867: warning: ignoring old recipe for target '../SuperRes/Obj/main.o'
Makefile:5295: warning: overriding recipe for target '../SuperRes/Obj/mainwindow.o'
Makefile:3994: warning: ignoring old recipe for target '../SuperRes/Obj/mainwindow.o'
/usr/local/cuda-9.1//bin/nvcc -m64 -O3 -c --compiler-options -fno-strict-aliasing -use_fast_math --ptxas-options=-v -DCMAKE_C_COMPILER=/usr/bin/gcc-5 -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -DCMAKE_C_COMPILER=/usr/bin/gcc-5 -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -I/usr/local/cuda-9.1/include/ -I/usr/local/cuda-9.1/include/  -L/usr/local/cuda-9.1/lib64/ -lcuda -lcudart -L/usr/local/cuda-9.1/lib64/ -lcudart -lcuda ../SuperRes/Filter.cu -o ../SuperRes/Obj/Filter_cuda.o 2>&1 | sed -r "s/\(([0-9]+)\)/:\1/g" 1>&2
In file included from /usr/local/cuda-9.1/include/host_config.h:50:0,
                 from /usr/local/cuda-9.1/include/cuda_runtime.h:78,
                 from <command-line>:0:
/usr/local/cuda-9.1/include/crt/host_config.h:121:2: error: #error -- unsupported GNU version! gcc versions later than 6 are not supported!
 #error -- unsupported GNU version! gcc versions later than 6 are not supported!
  ^~~~~
/usr/local/cuda-9.1//bin/nvcc -m64 -O3 -c --compiler-options -fno-strict-aliasing -use_fast_math --ptxas-options=-v -DCMAKE_C_COMPILER=/usr/bin/gcc-5 -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -DCMAKE_C_COMPILER=/usr/bin/gcc-5 -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -I/usr/local/cuda-9.1/include/ -I/usr/local/cuda-9.1/include/  -L/usr/local/cuda-9.1/lib64/ -lcuda -lcudart -L/usr/local/cuda-9.1/lib64/ -lcudart -lcuda ../SuperRes/SuperResolutionKernel.cu -o ../SuperRes/Obj/SuperResolutionKernel_cuda.o 2>&1 | sed -r "s/\(([0-9]+)\)/:\1/g" 1>&2
In file included from /usr/local/cuda-9.1/include/host_config.h:50:0,
                 from /usr/local/cuda-9.1/include/cuda_runtime.h:78,
                 from <command-line>:0:
/usr/local/cuda-9.1/include/crt/host_config.h:121:2: error: #error -- unsupported GNU version! gcc versions later than 6 are not supported!
 #error -- unsupported GNU version! gcc versions later than 6 are not supported!
  ^~~~~
/usr/local/cuda-9.1//bin/nvcc -D_DEBUG --use_fast_math -DCMAKE_C_COMPILER=/usr/bin/gcc-5 -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -I/usr/local/cuda-9.1/include/ -I/usr/local/cuda-9.1/include/  --machine 64 -c -o Filter_cuda.o ../SuperRes/Filter.cu
In file included from /usr/local/cuda-9.1/include/host_config.h:50:0,
                 from /usr/local/cuda-9.1/include/cuda_runtime.h:78,
                 from <command-line>:0:
/usr/local/cuda-9.1/include/crt/host_config.h:121:2: error: #error -- unsupported GNU version! gcc versions later than 6 are not supported!
 #error -- unsupported GNU version! gcc versions later than 6 are not supported!
  ^~~~~
Makefile:2212: recipe for target 'Filter_cuda.o' failed
make: *** [Filter_cuda.o] Error 1
12:06:17: The process "/usr/bin/make" exited with code 2.
Error while building/deploying project SuperRes (kit: Desktop Qt 5.11.0 GCC 64bit)
When executing step "Make"
12:06:17: Elapsed time: 00:00.

I have tried the following:

changing the compiler in Tools>Options>Kits as in this screenshot.
- I do get a warning saying

:-1: warning: "/usr/bin/gcc-6" is used by qmake, but "/usr/bin/x86_64-linux-gnu-gcc-5" is configured in the kit.
Please update your kit or choose a mkspec for qmake that matches your target environment better.

and likewise for G++ however as you can see it's just a different path. I actually went into the Qt mkspec files and changed it to gcc-5. I will go into detail about that next.

Changing the options in the Qt mkspec files.
- I went into the "linux-g++-64" and "linux-g++" mkspec folders (liux-g++ is the one it said it uses by default but I manually set the kit to use linux-g++-64) and followed the included files and changed all the options to gcc-5 and g++-5 respectively. You can see that it worked in the warning above.
Changing options in my Project.pro file
- I changed SuperRes.pro and here is the current contents.

#-------------------------------------------------
#
# Project created by QtCreator 2018-06-01T16:27:25
#
#-------------------------------------------------

QT       += core \
            network \
            gui \


greaterThan(QT_MAJOR_VERSION, 4): QT += widgets

TARGET = SuperRes
TEMPLATE = app

# The following define makes your compiler emit warnings if you use
# any feature of Qt which has been marked as deprecated (the exact warnings
# depend on your compiler). Please consult the documentation of the
# deprecated API in order to know how to port your code away from it.
DEFINES += QT_DEPRECATED_WARNINGS

# You can also make your code fail to compile if you use deprecated APIs.
# In order to do so, uncomment the following line.
# You can also select to disable deprecated APIs only up to a certain version of Qt.
#DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0x060000    # disables all the APIs deprecated before Qt 6.0.0


SOURCES += \
        main.cpp \
        mainwindow.cpp \
    FileThread.cpp \
    ImageFloatViewer.cpp \
    ImageIntViewer.cpp \
    ImgRead.cpp \
    main.cpp \
    MainViewer.cpp \
    mainwindow.cpp \
    StatusViewer.cpp \
    Stopwatch.cpp \
    SuperResolution.cpp \
    SuperResolutionThread.cpp \
    SuperResolutionViewer.cpp \
    UDPRead.cpp

HEADERS += \
        mainwindow.h \
    mainviewer.h \
    statusdata.h \
    cudaError.h \
    cudaMath.h \
    FileThread.h \
    Filter.h \
    ImageFloatViewer.h \
    ImageHeader.h \
    ImageIntViewer.h \
    ImgRead.h \
    mainviewer.h \
    MainViewer.h \
    mainwindow.h \
    statusdata.h \
    StatusData.h \
    StatusViewer.h \
    Stopwatch.h \
    SuperResolution.h \
    SuperResolutionKernel.h \
    SuperResolutionThread.h \
    SuperResolutionViewer.h \
    UDPRead.h \

FORMS += \
        mainwindow.ui \
    mainwindow.ui

DISTFILES += \
    SuperResViewer.suo \
    SuperRes.pro.user \
    SuperRes.pro.autosave \
    SuperResViewer.sln \
    SuperResViewer.vcxproj \
    SuperResViewer.vcxproj.filters \
    SuperResViewer.vcxproj.user \
    Filter.cu \
    SuperResolutionKernel.cu \
    Filter.cu \
    .cu

SUBDIRS += \
    SuperRes.pro

INCLUDEPATH += /usr/local/cuda-9.1/include/

LIBS += -L"/usr/local/cuda-9.1/lib64/" -lcuda -lcudart

#CONFIG += console
#CONFIG -= app_bundle

SYSTEM_TYPE = 64

DESTDIR = $$system(pwd)
CUDA_OBJECTS_DIR = ./
OBJECTS_DIR = $$DESTDIR/Obj


QMAKE_CXXFLAGS_RELEASE =-O3 -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5

CUDA_SOURCES += Filter.cu \
                SuperResolutionKernel.cu

# Path to cuda toolkit install
CUDA_DIR      = /usr/local/cuda-9.1/
# Path to header and libs files
INCLUDEPATH += /usr/local/cuda-9.1/include/
QMAKE_LIBDIR += /usr/local/cuda-9.1/lib64/    # Note I'm using a 64 bits Operating system
# libs used in your code
LIBS += -L"/usr/local/cuda-9.1/lib64/" -lcudart -lcuda
# GPU architecture
#CUDA_ARCH     = sm_21                # Yeah! I've a new device. Adjust with your compute capability
# Here are some NVCC flags I've always used by default.
NVCC_OPTIONS = --use_fast_math  -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5
NVCCFLAGS     = --compiler-options -fno-strict-aliasing -use_fast_math --ptxas-options=-v  -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5


# Prepare the extra compiler configuration (taken from the nvidia forum - i'm not an expert in this part)
CUDA_INC = $$join(INCLUDEPATH,' -I','-I',' ')
#-arch=$$CUDA_ARCH
cuda.commands = $$CUDA_DIR/bin/nvcc -m64 -O3  -c $$NVCCFLAGS  -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 \
                $$CUDA_INC $$LIBS  ${QMAKE_FILE_NAME} -o ${QMAKE_FILE_OUT} \
                2>&1 | sed -r \"s/\\(([0-9]+)\\)/:\\1/g\" 1>&2
# nvcc error printout format ever so slightly different from gcc
# http://forums.nvidia.com/index.php?showtopic=171651

cuda.dependency_type = TYPE_C # there was a typo here. Thanks workmate!
cuda.depend_command = $$CUDA_DIR/bin/nvcc -O3 -M $$CUDA_INC $$NVCCFLAGS  -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5   ${QMAKE_FILE_NAME}

cuda.input = CUDA_SOURCES
cuda.output = ${OBJECTS_DIR}${QMAKE_FILE_BASE}_cuda.o
# Tell Qt that we want add more stuff to the Makefile
QMAKE_EXTRA_COMPILERS += cuda

# SPECIFY THE R PATH FOR NVCC (this caused me a lot of trouble before)
QMAKE_LFLAGS += -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -Wl,-rpath,/usr/local/cuda-9.1/lib64/ # <-- added this

NVCCFLAGS =  -DCMAKE_C_COMPILER=/usr/bin/gcc-5  -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 -Xlinker -rpath,/usr/local/cuda-9.1/lib64/ # <-- and this
#-arch=$$CUDA_ARCH
CONFIG(debug, debug|release) {
    # Debug mode
    cuda_d.input = CUDA_SOURCES
    cuda_d.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
    cuda_d.commands = $$CUDA_DIR/bin/nvcc -D_DEBUG $$NVCC_OPTIONS $$CUDA_INC $$NVCC_LIBS --machine $$SYSTEM_TYPE  -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda_d.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda_d
}
else {
    # Release mode
    cuda.input = CUDA_SOURCES
    cuda.output = $$CUDA_OBJECTS_DIR/${QMAKE_FILE_BASE}_cuda.o
    cuda.commands = $$CUDA_DIR/bin/nvcc $$NVCC_OPTIONS $$CUDA_INC $$NVCC_LIBS --machine $$SYSTEM_TYPE -arch=$$CUDA_ARCH -c -o ${QMAKE_FILE_OUT} ${QMAKE_FILE_NAME}
    cuda.dependency_type = TYPE_C
    QMAKE_EXTRA_COMPILERS += cuda
}

I added -DCMAKE_CXX_COMPILER=/usr/bin/g++-5 and -DCMAKE_CXX_COMPILER=/usr/bin/gcc-5 wherever I could be it the NVCC flags or qmake options. I just put it wherever I could and it hasn't raised any extra warnings or errors so I can't imagine it's doing any harm but it's definitely not fixing the errors.

NOTE:

I was trying this with GCC/G++ 6 when I had 7 enabled first then I tried gcc/g++ 5 because I thought maybe it meant anything above 6.0 so 6.4.0 would also be bad. Again i do not know exactly what the cause of this is. I know it's coming from cuda not being compatible with GCC but I don't know if qt-creator might be what is setting the GCC version here. Thanks for your time.

NOTE2:

I wasn't sure what part of the forum to post this in. I thought General and Desktop might be appropriate.

SGaist · wrote on 6 Jun 2018, 19:49

Hi and welcome to devnet,

What Linux distribution are you running ?

rustyraptor · wrote on 6 Jun 2018, 19:55

Ok I just wasn't google searching the right phrases. This, at least, seems to fix the issue. Once I have successfully compiled the program I will delete this post or mark as a duplicate respecitvely.

https://stackoverflow.com/questions/14787148/cuda-and-gcc-compatibility-issue?rq=1

SGaist · 6 Jun 2018, 21:24

No reason to delete this post.

That will make the solution even more visible :)

rustyraptor · S SGaist 6 Jun 2018, 20:03

@SGaist alright will do.