[BUG] The case of iterate() + limit() seems not efficient

**Describe the bug**
A clear and concise description of what the bug is.

On personal mac, run the following the queries on LDBC graph with factor=1 with worker_num=32,
Query1：`g.V().hasLabel("PERSON").both("1..10").limit(100000)` costs 6s,
Query2: `g.V().hasLabel("PERSON").both("1..2").limit(100000)` costs 0.5s,
and `g.V().hasLabel("PERSON").both("1..2").count()` is 8366255.

In a sense, Query1 should not be slower than Query2 due to the early termination optimization. However, it is slower.