quiho: Automated Performance Regression Testing Using Inferred Resource Utilization Profil

quiho is a framework for profiling application performance to enable automated regression testing, introduced by Ivo Jimenez. It applies sensitivity analysis, specifically statistical regression analysis, using application-independent performance feature vectors to characterize machine performance. The resulting inferred resource utilization profiles serve as a proxy for hardware and low-level system software behavior to validate performance across code revisions.

Use Cases

• Automated performance regression testing based on inferred resource utilization profiles.
• Profiling application performance across code revisions using sensitivity analysis.
• Identifying hardware and low-level system software behavior based on feature importance from statistical regression analysis.

Strengths

• Framework is designed for automated testing without requiring code instrumentation or performance counters.
• Methodology is demonstrated on both synthetically introduced regressions and real-world applications.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

Ivo Jimenez

Collection Method

Framework for profiling application performance using statistical regression analysis on performance feature vectors.