One option would be to allow people to hoist the check manually and then elide the checks inside the loops.

func OnesCount(xs []uint64) int {
    var total int
    if cpu.HasSSE2 {
        for _, x := range xs {
            total += bits.OnesCount64(x)    
        }
    } else {
        for _, x := range xs {
            total += bits.OnesCount64(x)    
        }
    }
    return total
}

This becomes even more noticeable when the loop is unrolled at all, because then you don't even have another branch between the things.

I did some testing to find out what the impact was, and because I was not awake enough at the time, I ended up with an implementation that had the untaken branches to call the library functions, but not the popcount instructions, and discovered that the cost of that, plus the cost of the popcount operation without the branches, is much smaller than the cost of the branches and the popcount instructions. I'm not sure why. But the net impact is that the cost of the branch before every popcount, even though it's obviously a completely predictable branch (and perf confirms that it's predicted essentially 100% of the time), ends up being to increase the cost of popcount something like 3x overall.

@egonelbre One option would be to allow people to hoist the check manually

One issue with this is that internal/cpu is not a public API, and I am not sure it is something we want to surface to the user. /cc @martisch

The externally usable version of internal/cpu is golang.org/x/sys/cpu.

But, regardless, we should not expect users to manually hoist a CPU-specific flag.

This is fixed by fff7509

Change https://golang.org/cl/227238 mentions this issue: cmd/compile: use MOVBQZX for OpAMD64LoweredHasCPUFeature

I tried CL 227238. I'm using AMD Ryzen 5 3500U.

name      old time/op  new time/op  delta
FMA-8     1.16ns ± 0%  0.95ns ± 0%   ~     (p=1.000 n=1+1)
NonFMA-8  0.89ns ± 0%  0.98ns ± 0%   ~     (p=1.000 n=1+1)

name      old time/op  new time/op  delta
FMA-8     1.16ns ± 0%  0.73ns ± 0%   ~     (p=1.000 n=1+1)
NonFMA-8  0.89ns ± 0%  0.96ns ± 0%   ~     (p=1.000 n=1+1)

name      old time/op  new time/op  delta
FMA-8     1.16ns ± 0%  1.17ns ± 0%   ~     (p=1.000 n=1+1)
NonFMA-8  0.89ns ± 0%  1.21ns ± 0%   ~     (p=1.000 n=1+1)