Segmentation fault (core dumped) when using avx on an array allocated with new[]

Question

When I run this code in visual studio 2015, the code works correctly.But the code generates the following error in codeblocks : Segmentation fault(core dumped). I also ran the code in ubuntu with same error.

#include <iostream>
#include <immintrin.h>

struct INFO
{
    unsigned int id = 0;
    __m256i temp[8];
};

int main()
{
    std::cout<<"Start AVX..."<<std::endl;
    int _size = 100;
    INFO  *info = new INFO[_size];
    for (int i = 0; i<_size; i++)
    {
        for (int k = 0; k < 8; k++)
        {
            info[i].temp[k] = _mm256_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19,
                20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31);

        }
    }
    std::cout<<"End AVX."<<std::endl;
    return 0;
}

Answer 1

The problem is that prior to C++17 new and delete did not respect the alignment of the to-be-allocated type. If you look at the generated assembly from this simple function:

INFO* new_test() {
    int _size = 100;
    INFO  *info = new INFO[_size];
    return info;
}

You'll see that when compiled with anything prior to C++17 operator new[](unsigned long) is called, whereas for C++17 a call is made to operator new[](unsigned long, std::align_val_t) (and 32 is passed for the second parameter). Play around with it at godbolt.

If you can't use C++17, you can overwrite operator new[] (and operator delete[] -- and you should overwrite operator new and operator delete as well ...):

struct INFO {
    unsigned int id = 0;
    __m256i temp[8];
    void* operator new[](size_t size) {
        // part of C11:
        return aligned_alloc(alignof(INFO), size);
    }
    void operator delete[](void* addr) {
        free(addr); // aligned_alloc is compatible with free
    }
};

This is part of the previous godbolt example, if you compile with -DOVERWRITE_OPERATOR_NEW.

Note that this does not solve the alignment issue when using std::vector (or any other std-container), for that you need to pass an aligned allocator to the container (not part of the previous example).

Answer 2

I found two ways to solve this problem

The first solution How to solve the 32-byte-alignment issue for AVX load/store operations?

struct INFO
{
    __m256i temp[8];
    unsigned int id = 0;
};
INFO  *info = static_cast<INFO*>(_mm_malloc(sizeof(INFO)*_size, 32));
_mm_free(info);

The second solution

INFO  *info = new INFO[_size];
for (int i = 0; i < _size; i++)
{
    INFO new_info;
    for (int k = 0; k < 8; k++)
    {
        new_info.temp[k] = _mm256_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19,
            20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31);

    }
    info[i] = new_info;
}