perf: use bitmap::iter_ones() instead of bitmap::iter()

[st1page]

We have many logic like this

for ((a, b), vis) in aaa.iter()
.zip_eq(bbb.iter())
.zip_eq(chunk.visibility().iter()) {
   if !vis {
     continue
   }
   // do something
}

Use bitmap::iter_ones() can optimize the performance by skipping the invisible rows with SIMD bit operation #9278, especially when the selectivity visibility bitmap is low.

An example is #9329

[github-actions]

This issue has been open for 60 days with no activity. Could you please update the status? Feel free to continue discussion or close as not planned.