[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wenchen Fan updated SPARK-15690: Target Version/s: (was: 2.4.0) > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin >Priority: Major > > Spark's current shuffle implementation sorts all intermediate data by their > partition id, and then write the data to disk. This is not a big bottleneck > because the network throughput on commodity clusters tend to be low. However, > an increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > The goal of this ticket is to change Spark so it can use in-memory radix sort > to do data shuffling on a single node, and still gracefully fallback to disk > if the data size does not fit in memory. Given the number of partitions is > usually small (say less than 256), it'd require only a single pass do to the > radix sort with pretty decent CPU efficiency. > Note that there have been many in-memory shuffle attempts in the past. This > ticket has a smaller scope (single-process), and aims to actually > productionize this code. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sameer Agarwal updated SPARK-15690: --- Target Version/s: 2.4.0 (was: 2.3.0) > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin > > Spark's current shuffle implementation sorts all intermediate data by their > partition id, and then write the data to disk. This is not a big bottleneck > because the network throughput on commodity clusters tend to be low. However, > an increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > The goal of this ticket is to change Spark so it can use in-memory radix sort > to do data shuffling on a single node, and still gracefully fallback to disk > if the data size does not fit in memory. Given the number of partitions is > usually small (say less than 256), it'd require only a single pass do to the > radix sort with pretty decent CPU efficiency. > Note that there have been many in-memory shuffle attempts in the past. This > ticket has a smaller scope (single-process), and aims to actually > productionize this code. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Michael Armbrust updated SPARK-15690: - Target Version/s: 2.3.0 (was: 2.2.0) > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin > > Spark's current shuffle implementation sorts all intermediate data by their > partition id, and then write the data to disk. This is not a big bottleneck > because the network throughput on commodity clusters tend to be low. However, > an increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > The goal of this ticket is to change Spark so it can use in-memory radix sort > to do data shuffling on a single node, and still gracefully fallback to disk > if the data size does not fit in memory. Given the number of partitions is > usually small (say less than 256), it'd require only a single pass do to the > radix sort with pretty decent CPU efficiency. > Note that there have been many in-memory shuffle attempts in the past. This > ticket has a smaller scope (single-process), and aims to actually > productionize this code. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Reynold Xin updated SPARK-15690: Target Version/s: 2.2.0 (was: 2.1.0) > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin > > Spark's current shuffle implementation sorts all intermediate data by their > partition id, and then write the data to disk. This is not a big bottleneck > because the network throughput on commodity clusters tend to be low. However, > an increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > The goal of this ticket is to change Spark so it can use in-memory radix sort > to do data shuffling on a single node, and still gracefully fallback to disk > if the data size does not fit in memory. Given the number of partitions is > usually small (say less than 256), it'd require only a single pass do to the > radix sort with pretty decent CPU efficiency. > Note that there have been many in-memory shuffle attempts in the past. This > ticket has a smaller scope (single-process), and aims to actually > productionize this code. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Reynold Xin updated SPARK-15690: Description: Spark's current shuffle implementation sorts all intermediate data by their partition id, and then write the data to disk. This is not a big bottleneck because the network throughput on commodity clusters tend to be low. However, an increasing number of Spark users are using the system to process data on a single-node. When in a single node operating against intermediate data that fits in memory, the existing shuffle code path can become a big bottleneck. The goal of this ticket is to change Spark so it can use in-memory radix sort to do data shuffling on a single node, and still gracefully fallback to disk if the data size does not fit in memory. Given the number of partitions is usually small (say less than 256), it'd require only a single pass do to the radix sort with pretty decent CPU efficiency. Note that there have been many in-memory shuffle attempts in the past. This ticket has a smaller scope (single-process), and aims to actually productionize this code. was: An increasing number of Spark users are using the system to process data on a single-node. When in a single node operating against intermediate data that fits in memory, the existing shuffle code path can become a big bottleneck. Ideally, Spark should be able to use in-memory radix sort to do data shuffling on a single node > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin > > Spark's current shuffle implementation sorts all intermediate data by their > partition id, and then write the data to disk. This is not a big bottleneck > because the network throughput on commodity clusters tend to be low. However, > an increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > The goal of this ticket is to change Spark so it can use in-memory radix sort > to do data shuffling on a single node, and still gracefully fallback to disk > if the data size does not fit in memory. Given the number of partitions is > usually small (say less than 256), it'd require only a single pass do to the > radix sort with pretty decent CPU efficiency. > Note that there have been many in-memory shuffle attempts in the past. This > ticket has a smaller scope (single-process), and aims to actually > productionize this code. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-15690) Fast single-node (single-process) in-memory shuffle
[ https://issues.apache.org/jira/browse/SPARK-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Reynold Xin updated SPARK-15690: Summary: Fast single-node (single-process) in-memory shuffle (was: Fast single-node in-memory shuffle) > Fast single-node (single-process) in-memory shuffle > --- > > Key: SPARK-15690 > URL: https://issues.apache.org/jira/browse/SPARK-15690 > Project: Spark > Issue Type: New Feature > Components: Shuffle, SQL >Reporter: Reynold Xin > > An increasing number of Spark users are using the system to process data on a > single-node. When in a single node operating against intermediate data that > fits in memory, the existing shuffle code path can become a big bottleneck. > Ideally, Spark should be able to use in-memory radix sort to do data > shuffling on a single node -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org