systemml-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chamath Abeysinghe <abeysinghecham...@gmail.com>
Subject Re: Sub projects in Language and run time for parameter servers [SYSTEMML-2083]
Date Wed, 14 Mar 2018 15:11:45 GMT
Hi Matthias,
After going through JIRA sub projects and references you provide I thought
of drafting proposal focusing the Distributed spark backend
<https://issues.apache.org/jira/browse/SYSTEMML-2087> project because it
seems challenging and exciting area to explore :-).
I have sketched a rough diagram for design and the implementation plan for
the proposal,




My idea is making Paramserv runtime similar design to ParFor runtime, and
as a extension it will handle parameter exchange. So there I will work on
some primitives required by runtime to manage the PS and then in Spark I
will implement a parameter server. Initially it will work using synchronous
method and then if time allows I will experiment with other methods and
performance factors.

And also regarding the control program I have some concerns,
In the project JIRA it was mentioned that "PS strategies will be selected
by the user", does this include the architecture of the parameter server(#
of workers and servers)  also  or does it need to be handled in the
project?

I hope this plan aligns with expectations of the community and does not
conflicts with other GSoC candidates. Your feedback for this highly
appreciated, if there is anything wrong please correct me. Thanks

Regards,
Chamath


On Fri, Mar 9, 2018 at 2:19 PM, Matthias Boehm <mboehm7@gmail.com> wrote:

> Hi Chamath,
>
> ad 1: Yes, this is absolutely correct. However, it is important to realize
> that within the workers, we want to run dml functions, and for these we'll
> reuse our existing compiler, runtime, operations, and data structures.
>
> ad 2: Yes, this is also correct. Indeed we can use an existing parfor
> (with local execution mode) to emulate a local, synchronous parameter
> server. However, it would be very hard - and conflicting with our
> functional and thus, stateless execution semantics - to incorporate
> asynchronous updates and strategies such as Hogwild!. Furthermore, such a
> local parameter server might also have an application with very large
> models and batches, because this would enable distributed data-parallel
> operations spawn from each local worker.
>
> ad 3: Unfortunately, there is no one single detailed architecture diagram
> because the system evolves over time. I would recommend to look at the
> following two papers, where especially [1] (the parfor paper, and its
> extensions for Spark in [2]) might give you a better idea of the parameter
> server and its workers, which are primarily meant to handle the
> orchestration and efficient parameter updates/exchange. if you're looking
> for coarse-grained component, then [3], slide 8 might be a starting point.
> At a high-level each operation and some constructs like parfor have
> physical operators for CP, SPARK, MR, and some for GPU. Similarly this
> project aims to introduce a new paramserv builtin function (most similar to
> parfor) and its different physical operators.
>
> ad 4: Since this paramserv function has similarity with parfor, we will be
> able to reuse key primitives for bringing up local/remote workers, shipping
> the compiled functions, and input data. The major extensions will be to
> call the shipped functions per batch, get the returned (i.e., updated)
> parameters and handle the exchange accordingly to the paramserv
> configuration. However, since paramserv as an operation is implemented from
> scratch, we can customize as needed and are not restricted by script-level
> semantics which renders the problem simpler as the general-purpose parfor
> construct. Both have their use cases.
>
> In case this did not clarify your questions, let us known and we'll sort
> it out.
>
> [1] http://www.vldb.org/pvldb/vol7/p553-boehm.pdf, 2014
> [2] http://www.vldb.org/pvldb/vol9/p1425-boehm.pdf, 2016
> [3] http://boss.dima.tu-berlin.de/media/BOSS16-Tutorial-mboehm.pdf, 2016
>
> Regards,
> Matthias
>
> On Thu, Mar 8, 2018 at 10:28 PM, Chamath Abeysinghe <
> abeysinghechamath@gmail.com> wrote:
>
>> Hi,
>> I am trying to understand the purpose and work needed for different sub
>> projects in SYSTEMML-2083. And I got few questions,
>>
>> * In the JIRA it was mentioned that we are not integrating off the shelf
>> Parameter Server, but rather develop language and run time support from
>> scratch. As far as I understand, this means creating syntax for DML to
>> interact with the parameter server. And the parameter server implementation
>> is in different back-ends. So for example in Spark back end we have to
>> create a some kind of parameter server implementation with different
>> strategies, and it should be invoked by the syntax in DML. Is this
>> understanding correct?
>>
>> * In the JIRA there is a sub project for local multi threaded back-end.
>> In this project does "local" mean executing on single node similar to
>> ExecType.CP? If it is the case why use a parameter server for a single
>> node?
>>
>> * I was unable to find a architecture diagram for SystemML, is there any
>> that kind of diagram to understand the interaction between different
>> back-ends and language API or can you point me to those classes?
>>
>> * And those new run times, are they going to be completely new separate
>> run times or improvements to the existing ones?
>>
>> Please help me understand these issues. Thanks in advance.
>>
>> Regards,
>> Chamath
>>
>>
>


-- 
Chamath Abeysinghe
Department of Computer Science and Engineering
University of Moratuwa
Mobile: +94712803295

Mime
  • Unnamed multipart/mixed (inline, None, 0 bytes)
View raw message