Hello,
I’m working with the cuda 3.0 sdk, on windows XP.
The following source code makes cudafe segfault:
template <unsigned i>
struct Y
{
};
struct T
{
int m() { return 0; }
};
template <typename TT>
void test(Y<sizeof(&TT::m)>*)
{
}
int main()
{
test<T>(0);
return 0;
}
nvcc error : 'cudafe' died with status 0xC00000FD (STACK_OVERFLOW)
Will the c++ support improve in cuda 3.1?
Thanks.
–
Matthieu