CUDA 3.0 and gcc 4.4.x

Deadly combination! Compilation breaks when you try to do make on NVidia SDK example projects even. Problem is with nvcc incompatibility. A gcc version of 4.3.x works fine (as of this date).

Here is the compilation error listing:

/usr/include/string.h:43: error: inline function ‘void* memcpy(void*, const void*, size_t)’ cannot be declared weak

/usr/include/string.h:64: error: inline function ‘void* memset(void*, int, size_t)’ cannot be declared weak

/usr/include/bits/string3.h:49: error: inline function ‘void* memcpy(void*, const void*, size_t)’ cannot be declared weak

/usr/include/bits/string3.h:78: error: inline function ‘void* memset(void*, int, size_t)’ cannot be declared weak

/usr/local/cuda/bin/../include/common_functions.h:59: error: inline function ‘void* memset(void*, int, size_t)’ cannot be declared weak

/usr/local/cuda/bin/../include/common_functions.h:62: error: inline function ‘void* memcpy(void*, const void*, size_t)’ cannot be declared weak

/usr/local/cuda/bin/../include/math_functions.h:422: error: inline function ‘int __signbit(double)’ cannot be declared weak

/usr/local/cuda/bin/../include/math_functions.h:427: error: inline function ‘int __signbitf(float)’ cannot be declared weak

/usr/local/cuda/bin/../include/math_functions.h:440: error: inline function ‘int __signbitl(long double)’ cannot be declared weak

/usr/include/bits/mathcalls.h:350: error: inline function ‘int __signbit(double)’ cannot be declared weak

/usr/include/bits/mathcalls.h:350: error: inline function ‘int __signbitf(float)’ cannot be declared weak

/usr/include/bits/mathcalls.h:350: error: inline function ‘int __signbitl(long double)’ cannot be declared weak

/usr/include/bits/mathinline.h:38: error: inline function ‘int __signbitf(float)’ cannot be declared weak

/usr/include/bits/mathinline.h:50: error: inline function ‘int __signbit(double)’ cannot be declared weak

/usr/include/bits/mathinline.h:62: error: inline function ‘int __signbitl(long double)’ cannot be declared weak

/usr/include/c++/4.4/i486-linux-gnu/bits/c++locale.h: In function ‘int std::__convert_from_v(__locale_struct* const&, char*, int, const char*, ...)’:

/usr/include/c++/4.4/i486-linux-gnu/bits/c++locale.h:86: error: ‘__builtin_stdarg_start’ was not declared in this scope

dxtc.cu: In function ‘int main(int, char**)’:

dxtc.cu:789: warning: ignoring return value of ‘size_t fread(void*, size_t, size_t, FILE*)’, declared with attribute warn_unused_result

make[1]: *** [obj/i386/release/dxtc.cu.o] Error 1

Fixing the error basically means commenting out the specific lines in math_functions.h and common_functions.h on which the compiler shows the error (shown in dark blue).

Solution: comment them out. This will fix compilation errors. I have not tested the side-effect of this hack. But as for fixing compilation break, it works!

Note:

I needed a few more steps:

1. Uninstall gcc-4.4.x, g++-4.4.x (and all 4.4.x libraries)

2. Install gcc and g++ 4.3.x versions.

3. ln -s /usr/bin/gcc-4.3. /usr/bin/gcc

4. ln -s /usr/bin/g++-4.3. /usr/bin/g++

5. Install libXi-dev (since -lXi failed)

6. Install libxmu-dev (since -lxmu failed)

Steps 5,6 were required when I tried to build GL applications like fluidGL, simpleGL etc.

Related links:

http://www.piotrkrzyzek.com/nvidia-cuda-sdk-gcc-4-4-problems-solved-in-kubuntu-jaunt-910-usrbinld-cannot-find-lglut-and-__signbit__signbitf__signbitl__signbitd-cannot-be-declared-weak/

http://forums.nvidia.com/index.php?showtopic=83217

http://forums.nvidia.com/index.php?showtopic=87692