How to run, interpret, and compare FrameX benchmark results.

Performance Test

Use FrameX benchmark tooling to measure throughput, scaling, and memory behavior.

Run Full Benchmark Suite

python3 -m benchmarks.benchmark_suite

Fast Smoke Benchmark

python3 -m benchmarks.benchmark_suite \
  --rows 20000 \
  --array-elements 120000 \
  --object-items 30000 \
  --workers 1,2 \
  --repeats 1 \
  --warmups 0 \
  --skip-plots

Run Workload Capability Check

python3 -m benchmarks.check_framex_workloads

Output Artifacts

• benchmarks/results/benchmark_results.json
• benchmarks/results/benchmark_results.csv
• benchmarks/results/benchmark_report.md
• benchmarks/results/framex_workload_check.json
• plot images in benchmarks/results/*.png

Interpreting Key Metrics

• speedup > 1.0: FrameX faster than baseline
• seconds: lower is better
• peak_rss_mb: lower memory footprint is better

Optimization Workflow

1. Run benchmark suite and capture baseline artifacts.
2. Apply one optimization at a time.
3. Re-run with identical parameters.
4. Compare scenario-level speedup and memory impact.