[
https://issues.apache.org/jira/browse/SPARK-24374?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Xiangrui Meng updated SPARK-24374:
----------------------------------
Description:
(See details in the linked/attached SPIP doc.)
The proposal here is to add a new scheduling model to Apache Spark so users can
properly embed distributed DL training as a Spark stage to simplify the
distributed training workflow. For example, Horovod uses MPI to implement
all-reduce to accelerate distributed TensorFlow training. The computation model
is different from MapReduce used by Spark. In Spark, a task in a stage doesn’t
depend on any other tasks in the same stage, and hence it can be scheduled
independently. In MPI, all workers start at the same time and pass messages
around. To embed this workload in Spark, we need to introduce a new scheduling
model, tentatively named “barrier scheduling”, which launches tasks at the same
time and provides users enough information and tooling to embed distributed DL
training. Spark can also provide an extra layer of fault tolerance in case some
tasks failed in the middle, where Spark would abort all tasks and restart the
stage.
was:
(See details in the linked SPIP doc.)
The proposal here is to add a new scheduling model to Apache Spark so users can
properly embed distributed DL training as a Spark stage to simplify the
distributed training workflow. For example, Horovod uses MPI to implement
all-reduce to accelerate distributed TensorFlow training. The computation model
is different from MapReduce used by Spark. In Spark, a task in a stage doesn’t
depend on any other tasks in the same stage, and hence it can be scheduled
independently. In MPI, all workers start at the same time and pass messages
around. To embed this workload in Spark, we need to introduce a new scheduling
model, tentatively named “barrier scheduling”, which launches tasks at the same
time and provides users enough information and tooling to embed distributed DL
training. Spark can also provide an extra layer of fault tolerance in case some
tasks failed in the middle, where Spark would abort all tasks and restart the
stage.
> SPIP: Support Barrier Scheduling in Apache Spark
> ------------------------------------------------
>
> Key: SPARK-24374
> URL: https://issues.apache.org/jira/browse/SPARK-24374
> Project: Spark
> Issue Type: Story
> Components: ML, Spark Core
> Affects Versions: 3.0.0
> Reporter: Xiangrui Meng
> Assignee: Xiangrui Meng
> Priority: Major
> Labels: SPIP
> Attachments: SPIP_ Support Barrier Scheduling in Apache Spark.pdf
>
>
> (See details in the linked/attached SPIP doc.)
> The proposal here is to add a new scheduling model to Apache Spark so users
> can properly embed distributed DL training as a Spark stage to simplify the
> distributed training workflow. For example, Horovod uses MPI to implement
> all-reduce to accelerate distributed TensorFlow training. The computation
> model is different from MapReduce used by Spark. In Spark, a task in a stage
> doesn’t depend on any other tasks in the same stage, and hence it can be
> scheduled independently. In MPI, all workers start at the same time and pass
> messages around. To embed this workload in Spark, we need to introduce a new
> scheduling model, tentatively named “barrier scheduling”, which launches
> tasks at the same time and provides users enough information and tooling to
> embed distributed DL training. Spark can also provide an extra layer of fault
> tolerance in case some tasks failed in the middle, where Spark would abort
> all tasks and restart the stage.
--
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
