Vector Class Discussion

I want to perform a linear search through an integer array using SIMD instructions.

my idea is to iterate over the array and compare 4 integers in one call. My function should return the index of the matching element in the array. However, the Vec4i operator == does not return an index, only a vector of bools:

Vec4i a(10, 11, 12, 13);
Vec4i b(14, 13, 12, 11);
Vec4ib c = a == b;
// c = (false, false, true, false)

In order to get my index, I have to loop over the returned vector and check which element is true. Is there any way to avoid this overhead? i.e. the functions here (https://github.com/schani/linbin/blob/master/linear-sentinel-simd.h) use __builtin_ia32_pmovmskb128 and __builtin_ctz to calculate the index. And this paper here (http://openproceedings.org/EDBT/2014/paper_107.pdf) recommends popcount.

Thanks a lot,
Christoph