[jira] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16778660#comment-16778660 ] Xiangrui Meng commented on SPARK-24615: --- Attached the PDF files. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Xingbo Jiang >Priority: Major > Labels: Hydrogen, SPIP > Attachments: Accelerator-aware scheduling in Apache Spark 3.0.pdf, > SPIP_ Accelerator-aware scheduling.pdf > > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16778026#comment-16778026 ] Xingbo Jiang commented on SPARK-24615: -- I updated the SPIP and Product docs, please review and leave comments in this JIRA. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Xingbo Jiang >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16750872#comment-16750872 ] Saisai Shao commented on SPARK-24615: - I'm really sorry about the delay. Due to some changes on my side, I didn't have enough time to work on this before. But I talked to Xiangrui offline recently, we will continue to work on this and finalize it in 3.0. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16750751#comment-16750751 ] Felix Cheung commented on SPARK-24615: -- We are interested to know as well. [~mengxr] touched on this maybe Oct/Nov 2018, but I haven't heard anything else since. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16750375#comment-16750375 ] Thomas Graves commented on SPARK-24615: --- [~jerryshao] just curious where this is at, are you still working on it? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16567393#comment-16567393 ] Mingjie Tang commented on SPARK-24615: -- >From user's perspective, user only concern about the GPU resource for RDD, and >do not understand the stage or partitions of RDD. Therefore, underline >resource allocation mechanism would assign the resources to executor >automatically. Similar as cache or persistence to different level, maybe we can provide different configuration to users. Then, resource allocation to follow the predefined policy to allocate resource. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16564562#comment-16564562 ] Saisai Shao commented on SPARK-24615: - Leveraging dynamic allocation to tear down and bring up desired executor is a non-goal here, we will address it in feature, currently we're still focusing on using static allocation like spark.executor.gpus. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16564558#comment-16564558 ] Erik Erlandson commented on SPARK-24615: Am I understanding correctly that this can't assign executors to desired resources without resorting to Dynamic Allocation to tear down an Executor and reallocate it somewhere else? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16563730#comment-16563730 ] Saisai Shao commented on SPARK-24615: - Hi [~tgraves], I think eval() might unnecessarily break the lineage which could execute in one stage, for example: data transforming -> training -> transforming, this could possibly run in one stage, using eval will break into several stages, I'm not sure if it is the good choice. Also if we use eval to break the lineage, how do we store the intermediate data, like shuffle, or in memory/ on disk? Yes, how to break the boundaries is hard for user to know, but currently I cannot figure out a good solution, unless we use eval() to explicitly separate them. To solve the conflicts, failing might be one choice. Currently in my design, withResources only applies to the stage in which RDD will be executed, the following stages will still be ordinary stages without additional resouces. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16563711#comment-16563711 ] Thomas Graves commented on SPARK-24615: --- so I guess my question is this the right approach at all. Should we make it more obvious to the user where the boundaries would end by adding something like the eval(). I guess they could do like what they do for caching today which is force an action like count() to force the eval. If we are making the user add in their own action to evaluate, I would pick failing if there are conflicting resources. That way its completely obvious to the user that they might not be getting what they want. The only hard part is then though if the user doesn't know how to resolve that because of the way spark decides to pick its stages which might not be obvious to the users. Especially on the SQL side, but perhaps we aren't going to support that right now or support via hints? Its still not clear to me from the above are we saying the withResources applies to just the immediate next stage then, or it sticks until they change it again? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16555080#comment-16555080
]
Saisai Shao commented on SPARK-24615:
-
Hi [~tgraves] [~irashid] thanks a lot for your comments.
Currently in my design I don't insert a specific stage boundary with different
resources, the stage boundary is still the same (by shuffle or by result). so
{{withResouces}} is not an eval() action which trigger a stage. Instead, it
just adds a resource hint to the RDD.
So which means RDDs with different resources requirements in one stage may have
conflicts. For example: {{rdd1.withResources.mapPartitions \{ xxx
\}.withResources.mapPartitions \{ xxx \}.collect}}, resources in rdd1 may be
different from map rdd, so currently what I can think is that:
1. always pick the latter with warning log to say that multiple different
resources in one stage is illegal.
2. fail the stage with warning log to say that multiple different resources in
one stage is illegal.
3. merge conflicts with maximum resources needs. For example rdd1 requires 3
gpus per task, rdd2 requires 4 gpus per task, then the merged requirement would
be 4 gpus per task. (This is the high level description, details will be per
partition based merging) [chosen].
Take join for example, where rdd1 and rdd2 may have different resource
requirements, and joined RDD will potentially have other resource requirements.
For example:
{code}
val rddA = rdd.mapPartitions().withResources
val rddB = rdd.mapPartitions().withResources
val rddC = rddA.join(rddB).withResources
rddC.collect()
{code}
Here this 3 {{withResources}} may have different requirements. Since {{rddC}}
is running in a different stage, so there's no need to merge the resource
conflicts. But {{rddA}} and {{rddB}} are running in the same stage with
different tasks (partitions). So the merging strategy I'm thinking is based on
the partition, tasks running with partitions from {{rddA}} will use the
resource specified by {{rddA}}, so does {{rddB}}.
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16554742#comment-16554742
]
Imran Rashid commented on SPARK-24615:
--
hi, just catching up here -- agree with a lot of Tom's concerns. There are a
lot of corner cases we'd need to think about. Eg., you mentioned joins, that
in general they have stage boundaries -- but then again you might in some cases
have a shared partitioner, which would remove the stage boundary. And then
there are other cases like {{zip()}}.
What if you had two rdds in one stage with conflicting requirements? Eg. they
request different accelerator types, and no node in the cluster has both
available? Would you fail? Or introduce a stage boundary? I think failing is
fine, it would be OK if we require the user to put in their own boundary there
(eg. by writing to hdfs) just want to think it through.
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16554327#comment-16554327 ] Thomas Graves commented on SPARK-24615: --- Right so I think part of this is trying to make it more obvious to the user what the scope actually is. This in some ways is similar to caching. I see people many times force an evaluation after a cache to force the data to actually be cached because otherwise it might not do what they expect. that is why I mentioned the .eval() type functionality. For the example: val rddA = rdd.withResources.mapPartitions() val rddB = rdd.withResources.mapPartitions() val rddC = rddA.join(rddB) Above the mapPartitions would normally get their own stages correct? So I would think those stages would be with the resources specific but the join would be with the default resources. then you wouldn't have to worry about merging, etc. But you have the case with map or others where they wouldn't normally get their own stage so the question is perhaps should they, or do you provide something to? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16553679#comment-16553679 ] Saisai Shao commented on SPARK-24615: - [~Tagar] yes! > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16553121#comment-16553121
]
Ruslan Dautkhanov commented on SPARK-24615:
---
Is this work related to Spark's project Hydrogen's API?
[https://www.datanami.com/2018/06/05/project-hydrogen-unites-apache-spark-with-dl-frameworks/]
{quote}We also want to make Spark aware of accelerators so you can actually
comfortably use FPGA or GPUs in your latest clusters.
{quote}
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16552233#comment-16552233
]
Saisai Shao commented on SPARK-24615:
-
Hi [~tgraves] what you mentioned above is also what we think about and try to
figure out a way to solve it. (this problem also existed in barrier execution).
>From user point, specifying resource through RDD is the only feasible way
>currently what I can think, though resource is bound to stage/task not
>particular RDD. This means user could specify resources for different RDDs in
>a single stage, Spark can only use one resource within this stage. This will
>bring out several problems as you mentioned:
*Specify resources to which RDD*
For example {{rddA.withResource.mapPartition \{ xxx \}.collect()}} is not
different from {{rddA.mapPartition \{ xxx \}.withResource.collect}}. Since all
the rdds are executed in the same stage. So in the current design, not matter
the resource is specified with {{rddA}} or mapped RDD, the result is the same.
*one to one dependency RDDs with different resources*
For example {{rddA.withResource.mapPartition \{ xxx \}.withResource.collec()}},
here assuming the resource request for {{rddA}} and mapped RDD is different,
since they're running in a single stage, so we should fix such conflict.
*multiple dependencies RDDs with different resources*
For example:
{code}
val rddA = rdd.withResources.mapPartitions()
val rddB = rdd.withResources.mapPartitions()
val rddC = rddA.join(rddB)
{code}
If the resources in {{rddA}} is different from {{rddB}}, then we should also
fix such conflicts.
Previously I proposed to use largest resource requirement to satisfy all the
needs. But it may also cause the resource wasting, [~mengxr] mentioned to
set/merge resources per partition to avoid waste. In the meanwhile, it there's
a API exposed to set resources in the stage level, then this problem will not
be existed, but Spark doesn't expose such APIs to user, the only thing user can
specify is from RDD level, I'm still thinking of a good way to fix it.
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16551240#comment-16551240 ] Thomas Graves commented on SPARK-24615: --- the other thing which I think I mentioned above is could this handle saying, I want 1 node with 4GB, 4 cores and 3 nodes with 2 gpus, 10GB, 1cores) . > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16551233#comment-16551233
]
Thomas Graves commented on SPARK-24615:
---
Ok so thinking about this a bit more I slightly misread what it was applying
to. Really you are associated it with the new rdd that will be created (val
rddWithGPUResult = rdd.withResources.xxx) , not the original and to regenerate
the new rddwithGPUResult you would want to know it was created using those
resources. The thing that isn't clear to me is the scoping of this.
For instance if I say I have the code
val rdd1 = {{sc.textFile("README.md")}}
val rdd2 = rdd1.withResources.mapPartitions().collect().
Does the withResources apply to the entire line up to the action?
What if I change it to say
val rdd2 = rdd1.withResources.mapPartitions()
val res = rdd2.collect()
Does the withResources only apply to the mapPartitions, which is really what I
think you want with some of the ml algorithms. So we need to define what it
applies to. Doing something similar but would have more obvious scope:
val rdd2 = withResources {
rdd1.mapPartitions()
}
The above would be very obvious to the user the scope on it.
You also have things people could do like:
val rdd1 = rdd.withResources(x).mapPartitions()
val rdd2 = rdd.withResources(y).mapPartitions()
val rdd3 rdd1.join(rdd2)
I think in this case rdd1 and rdd2 have to be individually materialized before
you do the join for rdd3.
So its more like an implicit val rdd1 =
rdd.withResources(x).mapPartitions().eval() . You end up with putting in some
stage boundaries.
Have you thought about the scope and have ideas around this?
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16551198#comment-16551198 ] Mridul Muralidharan commented on SPARK-24615: - [~tgraves] This was indeed a recurring issue - the ability to modulate ask's to RM based on current requirements. What you bring out is an excellent point - changing resource requirements would be very useful - particularly for applications with heterogenous resource needs. Even currently when executor_memory/executor_cores does not allign well with stage requirements, we end up with OOM - resulting in over-provisioning memory needs; resulting in suboptimal use. GPU/accelerator aware scheduler is an extension of the same - where we have other resources to consider. I agree with [~tgraves] that a more general way to model this would look at all resources (when declaratively specified ofcourse) and use the information to allocate resources (from RM) and for task schedule (within spark). > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16550803#comment-16550803 ] Thomas Graves commented on SPARK-24615: --- yes if any requirement can't be satisfied it would use dynamic allocation to release and reacquire containers. I'm not saying we have to implement those parts right now, I'm saying we should keep them in mind during the design of this so they could be added later. I linked one old Jira that was about dynamically changing things. Its been brought up many times after in prs and just talking to customers not sure if there are other Jira as well. Its also somewhat related to SPARK-20589 where people just want to configure things per stage. I actually question if this should be done at the rdd level as well. A set of partitions don't care what the resources are, its generally the action you are taking on those rdd(s). Note it could be more then one rdd. I could do etl stuff on an RDD which resources would be totally different then if I ran tensorflow on that RDD for example. I do realize this is being tied in with the barrier stuff which is on the mapPartitions I'm not trying to be difficult and realize this Jira is more specific to the external ML algo's but don't want many api's for the same thing. I unfortunately haven't thought through a good solution for this, a while back my initial thought was to be able to pass in that resource context to the api calls, this obviously gets more tricky especially with pure sql support. I need to think about some more. the above proposal for .withResources is definitely closer but wonder about tying to the rdd still. cc [~irashid] [~mridulm80] who I think this has been brought up before with. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16550266#comment-16550266 ] Saisai Shao commented on SPARK-24615: - Hi [~tgraves] I'm still not sure how to handle memory per stage. Unlike MR, Spark shares the task runtime in a single JVM, I'm not sure how to control the memory usage within the JVM. Are you suggesting the similar approach like using GPU, when memory requirement cannot be satisfied, release the current executors and requesting new executors by dynamic resource allocation? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16550023#comment-16550023
]
Xiangrui Meng commented on SPARK-24615:
---
[~tgraves] Could you help link some past requests on configurable CPU/memory
per stage? And you are suggesting making the API generalizable to those
scenarios, but not including the feature under the scope of this proposal,
correct?
Btw, how do you like the following API?
{code:java}
rdd.withResources
.prefer("/gpu/k80", 2) // prefix of resource logical name, amount
.require("/cpu", 1)
.require("/memory", 819200)
.require("/disk", 1){code}
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16549345#comment-16549345 ] Thomas Graves commented on SPARK-24615: --- but my point is exactly that, it shouldn't be yet another mechanism for this, why not make a generic one that can handle all resources? > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16549276#comment-16549276
]
Saisai Shao commented on SPARK-24615:
-
Thanks [~tgraves] for the suggestion.
{quote}Once I get to the point I want to do the ML I want to ask for the gpu's
as well as ask for more memory during that stage because I didn't need more
before this stage for all the etl work. I realize you already have executors,
but ideally spark with the cluster manager could potentially release the
existing ones and ask for new ones with those requirements.
{quote}
Yes, I already discussed with my colleague offline, this is a valid scenario,
but I think to achieve this we should change the current dynamic resource
allocation mechanism Currently I marked this as a Non-Goal in this proposal,
only focus on statically resource requesting (--executor-cores,
--executor-gpus). I think we should support it later.
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16549266#comment-16549266 ] Thomas Graves commented on SPARK-24615: --- I think the usage for cpu/memory is the same. You know one job or stage has x number of tasks and perhaps they are caching at the time and need more memory or cpu. For instance look at your accelerator example. Let say I'm doing some etl before my ML. Once I get to the point I want to do the ML I want to ask for the gpu's as well as ask for more memory during that stage because I didn't need more before this stage for all the etl work. I realize you already have executors, but ideally spark with the cluster manager could potentially release the existing ones and ask for new ones with those requirements. That is obviously a separate task to do the latter part but I think if we are creating an interface we should keep those cases in mind. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16548662#comment-16548662 ] Saisai Shao commented on SPARK-24615: - Hi [~tgraves] I'm rewriting the design doc based on the comments mentioned above, so temporarily make it inaccessible, sorry about it, I will reopen it. I think it is hard to control the memory usage per stage/task, because task is running in the executor which shared within a JVM. For CPU, yes I think we can do it, but I'm not sure the usage scenario of it. For the requirement of using different types of machine, what I can think of is leveraging dynamic resource allocation. For example, if user wants run some MPI jobs with barrier enabled, then Spark could allocate some new executors with accelerator resource via cluster manager (for example using node label if it is running on YARN). But I will not target this as a goal in this design, since a) it is a non-goal for barrier scheduler currently; b) it makes the design too complex, would be better to separate to another work. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16547903#comment-16547903 ] Thomas Graves commented on SPARK-24615: --- did the design doc permissions change? I can't seem to access it now. A few overall concerns. We are now making accelerator configurations available per stage, but what about cpu and memory? It seems like if we are going to start making things configurable at the stage/rdd level it would be nice to be consistent. People have asked for this ability in the past. What about the case where to run some ML algorithm you would want machines of different types? For instance tensorflow with a parameter server might want gpu nodes for the workers but the parameter server would just be a cpu. This would also apply to the barrier scheduler so might cross post there. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16547263#comment-16547263 ] Saisai Shao commented on SPARK-24615: - Sure, I will also add it as Xiangrui also suggested the same concern. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16546617#comment-16546617 ] Thomas Graves commented on SPARK-24615: --- ok, I agree, I think this SPIP should at least cover how it could be integrated with the clusters managers to make sure its possible without large redesign later. The implementation could obviously wait. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16546590#comment-16546590
]
Saisai Shao commented on SPARK-24615:
-
Yes, currently the user is responsible for asking yarn to get resources like
GPU, for example like {{--num-gpus}}.
Yes, I agree with you, the concerns you mentioned above is valid. But currently
the design of this Jira only targets to the task level scheduling with
accelerator resources already available. It assumes that accelerator resources
is already got by executor and reported back to driver. Driver will schedule
the tasks based on the available resources.
For Spark to communicate with cluster manager to request other resources like
GPU, currently it is not covered in this design doc.
Xiangrui mentioned that Spark to communicate with cluster manager should also
be covered in this SPIP, so I'm still under drafting.
> Accelerator-aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Assignee: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16546580#comment-16546580 ] Thomas Graves commented on SPARK-24615: --- The user is responsible for asking yarn for the right executors? I assume perhaps based of the node label or something like that? Yarn in like 3.x has the ability to support other resource types like gpus, it would be much more efficient if you can ask the resource manager for the necessary resources, is there plan to support that? I also don't think its very good behavior to have the job just hang if user doesn't have the necessary resources. it feels like that should either be fail or if it can be linked to dynamic allocation that would be better. I left some questions in the design doc. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator-aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16545962#comment-16545962 ] Saisai Shao commented on SPARK-24615: - Sorry [~tgraves] for the late response. Yes, when requesting executors, user should know accelerators are required or not. If there's no satisfied accelerators, the job will be pending or not launched. > Accelerator-aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator aware task scheduling for Spark
[ https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16545253#comment-16545253 ] Thomas Graves commented on SPARK-24615: --- [~jerryshao] ^ > Accelerator aware task scheduling for Spark > --- > > Key: SPARK-24615 > URL: https://issues.apache.org/jira/browse/SPARK-24615 > Project: Spark > Issue Type: Improvement > Components: Spark Core >Affects Versions: 2.4.0 >Reporter: Saisai Shao >Assignee: Saisai Shao >Priority: Major > Labels: Hydrogen, SPIP > > In the machine learning area, accelerator card (GPU, FPGA, TPU) is > predominant compared to CPUs. To make the current Spark architecture to work > with accelerator cards, Spark itself should understand the existence of > accelerators and know how to schedule task onto the executors where > accelerators are equipped. > Current Spark’s scheduler schedules tasks based on the locality of the data > plus the available of CPUs. This will introduce some problems when scheduling > tasks with accelerators required. > # CPU cores are usually more than accelerators on one node, using CPU cores > to schedule accelerator required tasks will introduce the mismatch. > # In one cluster, we always assume that CPU is equipped in each node, but > this is not true of accelerator cards. > # The existence of heterogeneous tasks (accelerator required or not) > requires scheduler to schedule tasks with a smart way. > So here propose to improve the current scheduler to support heterogeneous > tasks (accelerator requires or not). This can be part of the work of Project > hydrogen. > Details is attached in google doc. It doesn't cover all the implementation > details, just highlight the parts should be changed. > > CC [~yanboliang] [~merlintang] -- 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] [Commented] (SPARK-24615) Accelerator aware task scheduling for Spark
[
https://issues.apache.org/jira/browse/SPARK-24615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16528293#comment-16528293
]
Thomas Graves commented on SPARK-24615:
---
maybe I"m missing it but how is this working with the resource manager?
rdd.map \{ }
.reduceByKey \{ xxx }
.mapPartition \{ accelerator required tasks }
.enableAccelerator()
.collect()
If I asked for 200 executors up front, it has to know this at time it asks for
them.
> Accelerator aware task scheduling for Spark
> ---
>
> Key: SPARK-24615
> URL: https://issues.apache.org/jira/browse/SPARK-24615
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
>Affects Versions: 2.4.0
>Reporter: Saisai Shao
>Priority: Major
> Labels: Hydrogen, SPIP
>
> In the machine learning area, accelerator card (GPU, FPGA, TPU) is
> predominant compared to CPUs. To make the current Spark architecture to work
> with accelerator cards, Spark itself should understand the existence of
> accelerators and know how to schedule task onto the executors where
> accelerators are equipped.
> Current Spark’s scheduler schedules tasks based on the locality of the data
> plus the available of CPUs. This will introduce some problems when scheduling
> tasks with accelerators required.
> # CPU cores are usually more than accelerators on one node, using CPU cores
> to schedule accelerator required tasks will introduce the mismatch.
> # In one cluster, we always assume that CPU is equipped in each node, but
> this is not true of accelerator cards.
> # The existence of heterogeneous tasks (accelerator required or not)
> requires scheduler to schedule tasks with a smart way.
> So here propose to improve the current scheduler to support heterogeneous
> tasks (accelerator requires or not). This can be part of the work of Project
> hydrogen.
> Details is attached in google doc. It doesn't cover all the implementation
> details, just highlight the parts should be changed.
>
> CC [~yanboliang] [~merlintang]
--
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
