diff --git a/integration_tests/src/main/scala/com/nvidia/spark/rapids/tests/common/BenchUtils.scala b/integration_tests/src/main/scala/com/nvidia/spark/rapids/tests/common/BenchUtils.scala
index 2bb87ef35d85..e9140b81a5a3 100644
--- a/integration_tests/src/main/scala/com/nvidia/spark/rapids/tests/common/BenchUtils.scala
+++ b/integration_tests/src/main/scala/com/nvidia/spark/rapids/tests/common/BenchUtils.scala
@@ -402,8 +402,8 @@ object BenchUtils {
       println(s"Both DataFrames contain $count1 rows")
 
       if (!ignoreOrdering && (df1.rdd.getNumPartitions > 1 || df2.rdd.getNumPartitions > 1)) {
-        throw new IllegalStateException(
-          "Cannot ignore ordering because one or more inputs have multiple partitions")
+        throw new IllegalStateException("Cannot run with ignoreOrdering=false because one or " +
+            "more inputs have multiple partitions")
       }
 
       val result1 = collectResults(df1, ignoreOrdering, useIterator)