Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-4590

Early investigation of parameter server

    XMLWordPrintableJSON

Details

    • Brainstorming
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • None
    • None
    • ML, MLlib
    • None

    Description

      In the currently implementation of GLM solvers, we save intermediate models on the driver node and update it through broadcast and aggregation. Even with torrent broadcast and tree aggregation added in 1.1, it is hard to go beyond ~10 million features. This JIRA is for investigating the parameter server approach, including algorithm, infrastructure, and dependencies.

      Attachments

        Issue Links

          Activity

            People

              rezazadeh Reza Zadeh
              mengxr Xiangrui Meng
              Votes:
              8 Vote for this issue
              Watchers:
              33 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: