[Breaking][jvm-packages] Bump rapids version to 22.12.0 (#8648)

* [jvm-packages] Bump rapids version to 22.12.0

This PR bumps spark version to 3.1.1 and the rapids version
to 22.12.0, which results in the latest xgboost can't run
with the old rapids packages.

jvm-packages/pom.xml

@@ -34,15 +34,15 @@
         <maven.compiler.source>1.8</maven.compiler.source>
         <maven.compiler.target>1.8</maven.compiler.target>
         <flink.version>1.8.3</flink.version>
-        <spark.version>3.0.3</spark.version>
+        <spark.version>3.1.1</spark.version>
         <scala.version>2.12.8</scala.version>
         <scala.binary.version>2.12</scala.binary.version>
         <hadoop.version>3.3.4</hadoop.version>
         <maven.wagon.http.retryHandler.count>5</maven.wagon.http.retryHandler.count>
         <log.capi.invocation>OFF</log.capi.invocation>
         <use.cuda>OFF</use.cuda>
-        <cudf.version>21.08.2</cudf.version>
+        <cudf.version>22.12.0</cudf.version>
-        <spark.rapids.version>21.08.0</spark.rapids.version>
+        <spark.rapids.version>22.12.0</spark.rapids.version>
         <cudf.classifier>cuda11</cudf.classifier>
     </properties>
     <repositories>

jvm-packages/xgboost4j-spark-gpu/src/main/scala/ml/dmlc/xgboost4j/scala/rapids/spark/GpuUtils.scala

@@ -17,7 +17,7 @@
 package ml.dmlc.xgboost4j.scala.rapids.spark
 
 import ai.rapids.cudf.Table
-import com.nvidia.spark.rapids.{ColumnarRdd, GpuColumnVector}
+import com.nvidia.spark.rapids.{ColumnarRdd, GpuColumnVectorUtils}
 import ml.dmlc.xgboost4j.scala.spark.util.Utils
 
 import org.apache.spark.rdd.RDD
@@ -31,16 +31,7 @@ private[spark] object GpuUtils {
 
   def extractBatchToHost(table: Table, types: Array[DataType]): Array[ColumnVector] = {
     // spark-rapids has shimmed the GpuColumnVector from 22.10
-    try {
+    GpuColumnVectorUtils.extractHostColumns(table, types)
-      val clazz = Utils.classForName("com.nvidia.spark.rapids.GpuColumnVectorUtils")
-      clazz.getDeclaredMethod("extractHostColumns", classOf[Table], classOf[Array[DataType]])
-        .invoke(null, table, types).asInstanceOf[Array[ColumnVector]]
-    } catch {
-      case _: ClassNotFoundException =>
-        // If it's older version, use the GpuColumnVector
-        GpuColumnVector.extractColumns(table, types).map(_.copyToHost())
-      case e: Throwable => throw e
-    }
   }
 
   def toColumnarRdd(df: DataFrame): RDD[Table] = ColumnarRdd(df)

jvm-packages/xgboost4j-spark-gpu/src/test/scala/ml/dmlc/xgboost4j/scala/rapids/spark/GpuTestSuite.scala

@@ -1,5 +1,5 @@
 /*
- Copyright (c) 2021 by Contributors
+ Copyright (c) 2021-2023 by Contributors
 
  Licensed under the Apache License, Version 2.0 (the "License");
  you may not use this file except in compliance with the License.
@@ -22,11 +22,10 @@ import java.util.{Locale, TimeZone}
 
 import org.scalatest.{BeforeAndAfterAll, FunSuite}
 
-import org.apache.spark.SparkConf
+import org.apache.spark.{GpuTestUtils, SparkConf}
 import org.apache.spark.internal.Logging
 import org.apache.spark.network.util.JavaUtils
 import org.apache.spark.sql.{Row, SparkSession}
-import org.apache.spark.sql.rapids.execution.TrampolineUtil
 
 trait GpuTestSuite extends FunSuite with TmpFolderSuite {
   import SparkSessionHolder.withSparkSession
@@ -232,7 +231,7 @@ object SparkSessionHolder extends Logging {
   }
 
   private def createSparkSession(): SparkSession = {
-    TrampolineUtil.cleanupAnyExistingSession()
+    GpuTestUtils.cleanupAnyExistingSession()
 
     // Timezone is fixed to UTC to allow timestamps to work by default
     TimeZone.setDefault(TimeZone.getTimeZone("UTC"))

jvm-packages/xgboost4j-spark-gpu/src/test/scala/org/apache/spark/GpuTestUtils.scala

@@ -0,0 +1,27 @@
+/*
+ Copyright (c) 2023 by Contributors
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ See the License for the specific language governing permissions and
+ limitations under the License.
+ */
+
+package org.apache.spark
+
+import org.apache.spark.sql.SparkSession
+
+object GpuTestUtils {
+
+  def cleanupAnyExistingSession(): Unit = {
+    SparkSession.cleanupAnyExistingSession()
+  }
+
+}

jvm-packages/xgboost4j-spark/src/main/scala/org/apache/spark/ml/util/XGBoostSchemaUtils.scala

@@ -1,5 +1,5 @@
 /*
- Copyright (c) 2022 by Contributors
+ Copyright (c) 2022-2023 by Contributors
 
  Licensed under the Apache License, Version 2.0 (the "License");
  you may not use this file except in compliance with the License.
@@ -18,7 +18,6 @@ package org.apache.spark.ml.util
 
 import org.apache.spark.sql.types.{BooleanType, DataType, NumericType, StructType}
 import org.apache.spark.ml.linalg.VectorUDT
-import org.apache.spark.ml.util.SchemaUtils
 
 object XGBoostSchemaUtils {
 

tests/buildkite/conftest.sh

@@ -24,7 +24,7 @@ set -x
 
 CUDA_VERSION=11.0.3
 RAPIDS_VERSION=22.10
-SPARK_VERSION=3.0.1
+SPARK_VERSION=3.1.1
 JDK_VERSION=8
 
 if [[ -z ${BUILDKITE:-} ]]