[VL] When performing a cast to int operation, an abnormal libgcc_s.so occurs

### Backend

VL (Velox)

### Bug description

When I execute
select count(1) from my_table where id > 0 (id in this table is string type),
scan and filter takes extra long (total execution time is about 400s in local mode). And abnormal libgcc.so appears in the flame graph.
But when I change to
select count(1) from my_table where id != '0'
or select count(1) from my_table where cast(id as double) >0,
the time is only about 150s. (Converting to double takes some extra time, but the flame graph is normal)

id != '0' 's flamegraph:


cast(id as double) >0 's flamegraph:


where cast (id as int) > 0 's flamegraph:


### Gluten version

Gluten-1.3

### Spark version

Spark-3.2.x

### Spark configurations

_No response_

### System information

--executor-memory 5g \
--driver-memory 3g \
--queue etron_dataz_new \
--conf spark.sql.legacy.allowNonEmptyLocationInCTAS=true \
--conf spark.memory.offHeap.enabled=true \
--conf spark.memory.offHeap.size=10g \
--conf spark.executor.memoryOverhead=2g \
--conf spark.shuffle.service.enabled=false \
--conf spark.sql.adaptive.enabled=true \
--conf spark.sql.optimizer.runtime.bloomFilter.applicationSideScanSizeThreshold=0 \
--conf spark.sql.optimizer.runtime.bloomFilter.enabled=true \
--conf spark.gluten.sql.columnar.physicalJoinOptimizeEnable=true \
--conf spark.gluten.sql.columnar.physicalJoinOptimizationLevel=18 \
--conf spark.gluten.sql.columnar.logicalJoinOptimizeEnable=true \
--conf spark.gluten.sql.columnar.logicalJoinOptimizationLevel=19 \

### Relevant logs

```bash

```