Documentation / Query Benchmark / Getting Started

Getting Started — Query Benchmark

Prefer Cross-Server mode when comparing hardware/versions with the same SQL text, and Version Compare when A/B testing two query rewrites on one server. Always pick a cache mode that matches whether you are measuring cold vs warm plans.

Getting Started

  1. Accept Terms and activate a license or start the 14-day trial.
  2. On 1. Setup, choose Single or Compare.
  3. Connect to SQL Server (and Query B server/database if needed).
  4. Paste SQL into Query A (and Query B for compare).
  5. Set iterations, concurrency, cache mode, and options.
  6. Click Run Benchmark. Review 2. Results, then save or export.

Setup Options

  • Concurrent users — Virtual clients running the query in parallel.
  • Iterations — Measured runs per user (or until duration expires).
  • Duration — Optional timed run (0 = use iteration count only).
  • Warm-up — Extra runs excluded from statistics (executed before cache flush).
  • MAXDOP — Optional OPTION (MAXDOP n) appended when not already present.
  • Isolation level — Optional SET TRANSACTION ISOLATION LEVEL … per session.
  • Collect actual plan — Capture plans after the run for viewing / comparison.
  • Live resource metrics — Instance counters while the run is active.

Licensing

14-day trial on first use; activate with a key from dblense.com. Manage license from About → License. Updates check the dblense-query-benchmark channel.