20424dad30
In `reduce` and `aggregate`, the driver node spends linear time on the number of partitions. It becomes a bottleneck when there are many partitions and the data from each partition is big. SPARK-1485 (#506) tracks the progress of implementing AllReduce on Spark. I did several implementations including butterfly, reduce + broadcast, and treeReduce + broadcast. treeReduce + BT broadcast seems to be right way to go for Spark. Using binary tree may introduce some overhead in communication, because the driver still need to coordinate on data shuffling. In my experiments, n -> sqrt(n) -> 1 gives the best performance in general, which is why I set "depth = 2" in MLlib algorithms. But it certainly needs more testing. I left `treeReduce` and `treeAggregate` public for easy testing. Some numbers from a test on 32-node m3.2xlarge cluster. code: ~~~ import breeze.linalg._ import org.apache.log4j._ Logger.getRootLogger.setLevel(Level.OFF) for (n <- Seq(1, 10, 100, 1000, 10000, 100000, 1000000)) { val vv = sc.parallelize(0 until 1024, 1024).map(i => DenseVector.zeros[Double](n)) var start = System.nanoTime(); vv.treeReduce(_ + _, 2); println((System.nanoTime() - start) / 1e9) start = System.nanoTime(); vv.reduce(_ + _); println((System.nanoTime() - start) / 1e9) } ~~~ out: | n | treeReduce(,2) | reduce | |---|---------------------|-----------| | 10 | 0.215538731 | 0.204206899 | | 100 | 0.278405907 | 0.205732582 | | 1000 | 0.208972182 | 0.214298272 | | 10000 | 0.194792071 | 0.349353687 | | 100000 | 0.347683285 | 6.086671892 | | 1000000 | 2.589350682 | 66.572906702 | CC: @pwendell This is clearly more scalable than the default implementation. My question is whether we should use this implementation in `reduce` and `aggregate` or put them as separate methods. The concern is that users may use `reduce` and `aggregate` as collect, where having multiple stages doesn't reduce the data size. However, in this case, `collect` is more appropriate. Author: Xiangrui Meng <meng@databricks.com> Closes #1110 from mengxr/tree and squashes the following commits: c6cd267 [Xiangrui Meng] make depth default to 2 b04b96a [Xiangrui Meng] address comments 9bcc5d3 [Xiangrui Meng] add depth for readability 7495681 [Xiangrui Meng] fix compile error 142a857 [Xiangrui Meng] merge master d58a087 [Xiangrui Meng] move treeReduce and treeAggregate to mllib 8a2a59c [Xiangrui Meng] Merge branch 'master' into tree be6a88a [Xiangrui Meng] use treeAggregate in mllib 0f94490 [Xiangrui Meng] add docs eb71c33 [Xiangrui Meng] add treeReduce fe42a5e [Xiangrui Meng] add treeAggregate |
||
|---|---|---|
| assembly | ||
| bagel | ||
| bin | ||
| conf | ||
| core | ||
| data/mllib | ||
| dev | ||
| docker | ||
| docs | ||
| ec2 | ||
| examples | ||
| external | ||
| extras | ||
| graphx | ||
| mllib | ||
| project | ||
| python | ||
| repl | ||
| sbin | ||
| sbt | ||
| sql | ||
| streaming | ||
| tools | ||
| yarn | ||
| .gitignore | ||
| .rat-excludes | ||
| .travis.yml | ||
| LICENSE | ||
| make-distribution.sh | ||
| NOTICE | ||
| pom.xml | ||
| README.md | ||
| scalastyle-config.xml | ||
| tox.ini | ||
Apache Spark
Spark is a fast and general cluster computing system for Big Data. It provides high-level APIs in Scala, Java, and Python, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and structured data processing, MLLib for machine learning, GraphX for graph processing, and Spark Streaming.
Online Documentation
You can find the latest Spark documentation, including a programming guide, on the project webpage at http://spark.apache.org/documentation.html. This README file only contains basic setup instructions.
Building Spark
Spark is built on Scala 2.10. To build Spark and its example programs, run:
./sbt/sbt assembly
(You do not need to do this if you downloaded a pre-built package.)
Interactive Scala Shell
The easiest way to start using Spark is through the Scala shell:
./bin/spark-shell
Try the following command, which should return 1000:
scala> sc.parallelize(1 to 1000).count()
Interactive Python Shell
Alternatively, if you prefer Python, you can use the Python shell:
./bin/pyspark
And run the following command, which should also return 1000:
>>> sc.parallelize(range(1000)).count()
Example Programs
Spark also comes with several sample programs in the examples directory.
To run one of them, use ./bin/run-example <class> [params]. For example:
./bin/run-example SparkPi
will run the Pi example locally.
You can set the MASTER environment variable when running examples to submit
examples to a cluster. This can be a mesos:// or spark:// URL,
"yarn-cluster" or "yarn-client" to run on YARN, and "local" to run
locally with one thread, or "local[N]" to run locally with N threads. You
can also use an abbreviated class name if the class is in the examples
package. For instance:
MASTER=spark://host:7077 ./bin/run-example SparkPi
Many of the example programs print usage help if no params are given.
Running Tests
Testing first requires building Spark. Once Spark is built, tests can be run using:
./sbt/sbt test
A Note About Hadoop Versions
Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported
storage systems. Because the protocols have changed in different versions of
Hadoop, you must build Spark against the same version that your cluster runs.
You can change the version by setting -Dhadoop.version when building Spark.
For Apache Hadoop versions 1.x, Cloudera CDH MRv1, and other Hadoop versions without YARN, use:
# Apache Hadoop 1.2.1
$ sbt/sbt -Dhadoop.version=1.2.1 assembly
# Cloudera CDH 4.2.0 with MapReduce v1
$ sbt/sbt -Dhadoop.version=2.0.0-mr1-cdh4.2.0 assembly
For Apache Hadoop 2.2.X, 2.1.X, 2.0.X, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions
with YARN, also set -Pyarn:
# Apache Hadoop 2.0.5-alpha
$ sbt/sbt -Dhadoop.version=2.0.5-alpha -Pyarn assembly
# Cloudera CDH 4.2.0 with MapReduce v2
$ sbt/sbt -Dhadoop.version=2.0.0-cdh4.2.0 -Pyarn assembly
# Apache Hadoop 2.2.X and newer
$ sbt/sbt -Dhadoop.version=2.2.0 -Pyarn assembly
When developing a Spark application, specify the Hadoop version by adding the
"hadoop-client" artifact to your project's dependencies. For example, if you're
using Hadoop 1.2.1 and build your application using SBT, add this entry to
libraryDependencies:
"org.apache.hadoop" % "hadoop-client" % "1.2.1"
If your project is built with Maven, add this to your POM file's <dependencies> section:
<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-client</artifactId>
<version>1.2.1</version>
</dependency>
Configuration
Please refer to the Configuration guide in the online documentation for an overview on how to configure Spark.
Contributing to Spark
Contributions via GitHub pull requests are gladly accepted from their original author. Along with any pull requests, please state that the contribution is your original work and that you license the work to the project under the project's open source license. Whether or not you state this explicitly, by submitting any copyrighted material via pull request, email, or other means you agree to license the material under the project's open source license and warrant that you have the legal authority to do so.