The compiler automatically peeled iterations from the vector loop into a scalar loop to align the vector loop with a particular memory reference; however, this optimization may not be ideal. To possibly achieve better performance, disable automatic peel generation using the directive: #pragma vector nodynamic_align

Example (original code)

...
#pragma vector nodynamic_align
for (int i = 0; i < len; i++)
...

void f(float * a, float * b, float * c, int len)
{
    #pragma vector nodynamic_align
    for (int i = 0; i < len; i++)
    {
        a[i] = b[i] * c[i];
    }
}

Read More

• vector
• Dynamic Alignment Clause for Vector Pragma/Directive and Vectorization Resources for Intel® Advisor Users