200ms regression per query when using optimizationTimeNanos compared to using optimizationTimeMillis

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Won't Fix
    • Priority: Major - P3
    • 9.0.0-rc0
    • Affects Version/s: None
    • Component/s: None
    • Query Optimization
    • ALL
    • Hide

      Run Synthetic benchmarks with 

       

      "internalMeasureQueryExecutionTimeInNanoseconds": true

       

      Show
      Run Synthetic benchmarks with    "internalMeasureQueryExecutionTimeInNanoseconds" : true  
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      While developing https://github.com/10gen/dsi/pull/4315/ It was observed that there's a 200ms regression per query when using optimizationTimeNanos compared to using optimizationTimeMillis,

       

      and was confirmed by running the same query set with and without Nano precision in a 3-time run patch at: https://performance-analyzer.server-tig.prod.corp.mongodb.com/perf-analyzer-viz/?comparison_id=69b82f4ea9505d426a53802e&selected_tab=scatter-plots&measurement_filter=AverageLatency&measurement_filter=OperationThroughput&percent_filter=0%7C%7C100&z_filter=0%7C%7C10&filter_type=Default 

       

        1. Screenshot 2026-03-16 at 15.57.36.png
          190 kB
          Hagar Mohamed
        2. Screenshot 2026-03-16 at 19.46.22.png
          154 kB
          Hagar Mohamed

            Assignee:
            Alexander Ignatyev
            Reporter:
            Hagar Mohamed
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: