store operations

void linear(int m, int n, int k, const float* A, int lda, const float* B, int ldb, float* scores, int lds, float* bias)
{
    int num_threads = n/32;
    omp_set_num_threads(num_threads);
#pragma omp parallel
    {
        __m256 b0 = _mm256_load_ps(&bias[32*t + 0]);
        __m256 b1 = _mm256_load_ps(&bias[32 * t + 8]);
        __m256 b2 = _mm256_load_ps(&bias[32 * t + 16]);
        __m256 b3 = _mm256_load_ps(&bias[32 * t + 24]);
        for (int i = 0; i < m; i++) {
            int storeOffset = i * lds + t * slice_rows;
            __m256 intSum[4] = { b0, b1, b2, b3 };
            for (int j = 0; j < k-16; j += 16) {
                //standard FMA operations for multiplication of matrix
            }
            _mm256_stream_ps(scores + storeOffset, intSum[0]);
            _mm256_stream_ps(scores + storeOffset + 8, intSum[1]);
            _mm256_stream_ps(scores + storeOffset + 16, intSum[2]);
            _mm256_stream_ps(scores + storeOffset + 24, intSum[3]);
        }
    }
}