Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-5384

Perf regression on basic standard test throughput

      On the December-17 YCSB performance build we saw an 8% regression (29056 vs 26712) compared to the previous December-15 build for the "Basic Standard Test Throughput" metric, against the WiredTiger develop branch on Linux. On Windows it had an even sharper 33% regression (22665 vs 15076) for the same metric.

      The list of WT tickets merged between the December-15 and December-17 builds are:

      1. Revert "WT-4968 remove all committed and connection level commit timestamps (#5040)" (#5053) (commit: 2b7cf1e) (detail / githubweb)
      2. WT-5314 For python tests, don't try to load extensions with empty names (#5049) (commit: b41d10a) (detail / githubweb)
      3. WT-5257 add assert to hopefully silence coverity (and for actual errors) (#5046) (commit: 2d2e285) (detail / githubweb)
      4. WT-5139 Incremental backup API (#4993) (commit: 77f9aeb) (detail / githubweb)

       

            Assignee:
            backlog-server-storage-engines [DO NOT USE] Backlog - Storage Engines Team
            Reporter:
            luke.chen@mongodb.com Luke Chen
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

              Created:
              Updated: