[[["이해하기 쉬움","easyToUnderstand","thumb-up"],["문제가 해결됨","solvedMyProblem","thumb-up"],["기타","otherUp","thumb-up"]],[["이해하기 어려움","hardToUnderstand","thumb-down"],["잘못된 정보 또는 샘플 코드","incorrectInformationOrSampleCode","thumb-down"],["필요한 정보/샘플이 없음","missingTheInformationSamplesINeed","thumb-down"],["번역 문제","translationIssue","thumb-down"],["기타","otherDown","thumb-down"]],["최종 업데이트: 2025-09-04(UTC)"],[[["\u003cp\u003eThe latest Dataproc Serverless runtime update notices are available in the Dataproc Serverless release notes.\u003c/p\u003e\n"],["\u003cp\u003eSpark runtime version 2.2 utilizes Apache Spark 3.5.1, Cloud Storage Connector 3.0.3, BigQuery Connector 0.36.4, Java 17, Conda 24.1, Python 3.12, R 4.3, and Scala 2.13 across all listed releases.\u003c/p\u003e\n"],["\u003cp\u003eDataproc Serverless includes machine learning libraries like TensorFlow, PyTorch, and XGBoost, as well as GPU-specific libraries like Spark Rapids 24.04.0 and NVIDIA Driver 550.127.05, to enhance machine learning and data science tasks.\u003c/p\u003e\n"],["\u003cp\u003eA variety of Python and R libraries are available in Dataproc Serverless for Spark runtime version 2.2, facilitating a broad range of data science and machine learning applications, with the complete lists of package names and versions provided.\u003c/p\u003e\n"],["\u003cp\u003eUsing distributed Spark XGBoost on a Dataproc Serverless batch workload with autoscaling enabled is not supported; dynamic allocation needs to be disabled to utilize XGBoost in such a scenario.\u003c/p\u003e\n"]]],[],null,["# Serverless for Apache Spark Spark runtime 2.2.x\n\n| See the [Serverless for Apache Spark release notes](/dataproc-serverless/docs/release-notes) for the latest Serverless for Apache Spark runtime update notices.\n\nSpark runtime version 2.2 components\n------------------------------------\n\n**Notes:**\n\n\n1. The `2.2` runtime uses the `UTF-8` default character encoding.\n\nSpark runtime 2.2 libraries\n---------------------------\n\nlearning libraries, such [TensorFlow](https://www.tensorflow.org/),\n[PyTorch](https://pytorch.org/), and [XGBoost](https://www.nvidia.com/en-us/glossary/xgboost/),\nand offer a ready-to-use environment for machine learning and data science\napplications.\n\nThe following sections list the library versions that are available in\nServerless for Apache Spark runtime version `2.2`.\n\n### GPU-specific libraries\n\nFor Serverless for Apache Spark batch workloads that use GPU VMs,\nthe following NVIDIA driver and libraries are available in the\nServerless for Apache Spark container. You can use them to accomplish the following\ntasks:\n\n- Accelerate Spark batch workloads with the [NVIDIA Spark Rapids library](https://docs.nvidia.com/spark-rapids/index.html)\n- Train machine learning workloads\n- Run distributed batch inference using Spark\n\n### XGBoost libraries\n\nThe following [Maven package versions](https://mvnrepository.com/artifact/ml.dmlc)\nare available in Serverless for Apache Spark runtime version `2.2` to use\n[XGBoost](https://www.nvidia.com/en-us/glossary/xgboost/) with Spark in Java or Scala.\n\n| **Note:** You cannot use distributed Spark XGBoost on a Serverless for Apache Spark batch workload that has [autoscaling](/dataproc-serverless/docs/concepts/autoscaling) enabled (the default behavior) because new nodes that start elastic scaling cannot receive new tasks and remain idle. To use XGBoost with a batch workload, you can set the [`spark.dynamicAllocation.enabled = false`](/dataproc-serverless/docs/concepts/autoscaling#spark_dynamic_allocation_properties) property on a batch workload to disable dynamic allocation.\n\n### Python libraries\n\nThe following Python library versions are included in\nServerless for Apache Spark runtime version `2.2`.\nserverless-spark-2.2-debian-12 python libraries\n\n### R libraries\n\nThe following R library versions are included in\nServerless for Apache Spark runtime version `2.2`.\nserverless-spark-2.2-debian-12 r libraries"]]
