[ https://issues.apache.org/jira/browse/GIRAPH-259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13426702#comment-13426702 ]
Alessandro Presta commented on GIRAPH-259: ------------------------------------------ Is there any particular reason why MasterCompute#compute() is executed before anything else? It seems to me that we would be more loyal to the BSP model if we didn't enforce this: MasterCompute is just a special vertex, executed during the superstep, whose vote to halt wins over everyone else's. You still use it as the logical place where serial, graph-wide computations are done (for example, checking the global value of an aggregator as a termination condition), but you don't rely on the contract that it will be run at any specific time relative to normal vertices. Especially since we are going to offer auto-resetting aggregators, it makes sense for MasterCompute to be a normal aggregator user, so collecting aggregators and running compute() are two orthogonal components. @Avery: regarding those two failures, please see GIRAPH-276 > TestBspBasic.testBspPageRank is broken > -------------------------------------- > > Key: GIRAPH-259 > URL: https://issues.apache.org/jira/browse/GIRAPH-259 > Project: Giraph > Issue Type: Bug > Reporter: Maja Kabiljo > Assignee: Maja Kabiljo > Attachments: GIRAPH-259-1.patch, GIRAPH-259-2.patch, > GIRAPH-259-3.patch > > > Test crashes on line 152 in class SimplePageRankVertex in distributed mode. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira