[VL] inputIteratorTransformer exhibits unexpectedly high time cost

### Backend

VL (Velox)

### Bug description

Currently, while executing vectorized etls, we unexpectedly found that inputIteratorTransformer is consuming a significant amount of CPU time.

The Spark UI is as follows:

The image shows the process from Table Scan to inputIteratorTransformer.
inputIteratorTransformer includes the total execution time of all previous operators, but from the CPU time perspective, it doesn’t seem to be the case.

Spark version: spark-3.0
Gluten version: branch-1.2

### Spark version

None

### Spark configurations

_No response_

### System information

_No response_

### Relevant logs

_No response_