[
https://issues.apache.org/jira/browse/SPARK-3623?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Guoqiang Li updated SPARK-3623:
-------------------------------
Description:
Consider the following code:
{code}
for (i <- 0 until totalIter) {
val previousCorpus = corpus
logInfo("Start Gibbs sampling (Iteration %d/%d)".format(i, totalIter))
val corpusTopicDist = collectTermTopicDist(corpus, globalTopicCounter,
sumTerms,
numTerms, numTopics, alpha, beta).persist(storageLevel)
val corpusSampleTopics = sampleTopics(corpusTopicDist,
globalTopicCounter, sumTerms, numTerms,
numTopics, alpha, beta).persist(storageLevel)
corpus = updateCounter(corpusSampleTopics,
numTopics).persist(storageLevel)
globalTopicCounter = collectGlobalCounter(corpus, numTopics)
assert(bsum(globalTopicCounter) == sumTerms)
previousCorpus.unpersistVertices()
corpusTopicDist.unpersistVertices()
corpusSampleTopics.unpersistVertices()
}
{code}
If there is no checkpoint operation will appear the following problems.
1. The RDD of corpus dependencies are too deep
2. The shuffle files are too large.
3. Any of a server crash will cause the algorithm to recalculate
was:
Consider the following code:
{code}
for (i <- 0 until totalIter) {
val previousCorpus = corpus
logInfo("Start Gibbs sampling (Iteration %d/%d)".format(i, totalIter))
val corpusTopicDist = collectTermTopicDist(corpus, globalTopicCounter,
sumTerms,
numTerms, numTopics, alpha, beta).persist(storageLevel)
val corpusSampleTopics = sampleTopics(corpusTopicDist,
globalTopicCounter, sumTerms, numTerms,
numTopics, alpha, beta).persist(storageLevel)
corpus = updateCounter(corpusSampleTopics,
numTopics).persist(storageLevel)
globalTopicCounter = collectGlobalCounter(corpus, numTopics)
assert(bsum(globalTopicCounter) == sumTerms)
previousCorpus.unpersistVertices()
corpusTopicDist.unpersistVertices()
corpusSampleTopics.unpersistVertices()
}
{code}
If there is no checkpoint operation will appear the following problems.
1. The RDD of corpus dependencies are too deep
2. The shuffle files are too large.
> GraphX does not support the checkpoint operation
> ------------------------------------------------
>
> Key: SPARK-3623
> URL: https://issues.apache.org/jira/browse/SPARK-3623
> Project: Spark
> Issue Type: Improvement
> Components: GraphX
> Affects Versions: 1.0.2, 1.1.0
> Reporter: Guoqiang Li
> Priority: Critical
>
> Consider the following code:
> {code}
> for (i <- 0 until totalIter) {
> val previousCorpus = corpus
> logInfo("Start Gibbs sampling (Iteration %d/%d)".format(i, totalIter))
> val corpusTopicDist = collectTermTopicDist(corpus, globalTopicCounter,
> sumTerms,
> numTerms, numTopics, alpha, beta).persist(storageLevel)
> val corpusSampleTopics = sampleTopics(corpusTopicDist,
> globalTopicCounter, sumTerms, numTerms,
> numTopics, alpha, beta).persist(storageLevel)
> corpus = updateCounter(corpusSampleTopics,
> numTopics).persist(storageLevel)
> globalTopicCounter = collectGlobalCounter(corpus, numTopics)
> assert(bsum(globalTopicCounter) == sumTerms)
> previousCorpus.unpersistVertices()
> corpusTopicDist.unpersistVertices()
> corpusSampleTopics.unpersistVertices()
> }
> {code}
> If there is no checkpoint operation will appear the following problems.
> 1. The RDD of corpus dependencies are too deep
> 2. The shuffle files are too large.
> 3. Any of a server crash will cause the algorithm to recalculate
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
