Performance and soak testing

The automated regression target exercises ten curve buffers with 10,000 samples each, a large render range, and repeated window creation and destruction:

make test-qtstriptool-performance

Its time limits are broad regression guards, not production capacity claims. For release acceptance, compare Qt and Motif on the same host and .stp file, with the same PV workload, window size, style, and Channel Access settings. Record CPU and memory, update delivery, repaint latency, reconnects, large archive requests, startup/shutdown, and a long run without unbounded memory growth.

The performance and soak procedure gives the full measurement plan and known evidence. Site owners should set numeric acceptance thresholds before measuring a release candidate.