crunch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gabriel Reid (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CRUNCH-294) Cost-based job planning
Date Sat, 16 Nov 2013 18:49:21 GMT

    [ https://issues.apache.org/jira/browse/CRUNCH-294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13824566#comment-13824566
] 

Gabriel Reid commented on CRUNCH-294:
-------------------------------------

Good point about the original issue that was brought up, i.e. minimizing CPU costs.

I'm thinking that having something like cpuCost() and memoryCost() methods on DoFn might be
the easiest/most flexible for now, with both methods returning 1.0f as the default value.
For now we could treat any cpuCost above 1.0f as "attempt to run only once", but using costs
instead of a flag will probably allow more flexibility in the future if we want to do more
advanced stuff in the planner.

That actually makes me think of something else to consider: whether cpuCost should be something
that should be interpreted independently, or should it be interpreted as scaleFactor * cpuCost?

> Cost-based job planning
> -----------------------
>
>                 Key: CRUNCH-294
>                 URL: https://issues.apache.org/jira/browse/CRUNCH-294
>             Project: Crunch
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Josh Wills
>            Assignee: Josh Wills
>         Attachments: CRUNCH-294.patch, jobplan-default-new.png, jobplan-default-old.png,
jobplan-large_s2_s3.png, jobplan-lopsided.png
>
>
> A bug report on the user list drove me to revisit some of the core planning logic, particularly
around how we decide where to split up DoFns between two dependent MapReduce jobs.
> I found an old TODO about using the scale factor from a DoFn to decide where to split
up the nodes between dependent GBKs, so I implemented a new version of the split algorithm
that takes advantage of how we've propagated support for multiple outputs on both the map
and reduce sides of a job to do finer-grained splits that use information from the scaleFactor
calculations to make smarter split decisions.
> One high-level change along with this: I changed the default scaleFactor() value in DoFn
to 0.99f to slightly prefer writes that occur later in a pipeline flow by default.



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message