Abstract:Software systems often have numerous configuration options that can be adjusted to meet different performance requirements. However, understanding the combined impact of these options on performance is often challenging, especially with limited real-world data. To tackle this issue, deep learning techniques have gained popularity due to their ability to capture complex relationships even with limited samples. This thesis begins with a systematic literature review of deep learning techniques in configuration performance modeling, analyzing 85 primary papers out of 948 searched papers. It identifies knowledge gaps and sets three objectives for the thesis. The first knowledge gap is the lack of understanding about which encoding scheme is better and in what circumstances. To address this, the thesis conducts an empirical study comparing three popular encoding schemes. Actionable suggestions are provided to support more reliable decisions. Another knowledge gap is the sparsity inherited from the configuration landscape. To handle this, the thesis proposes a model-agnostic and sparsity-robust framework called DaL, which uses a "divide-and-learn" approach. DaL outperforms state-of-the-art approaches in accuracy improvement across various real-world systems. The thesis also addresses the limitation of predicting under static environments by proposing a sequential meta-learning framework called SeMPL. Unlike traditional meta-learning frameworks, SeMPL trains meta-environments in a specialized order, resulting in significantly improved prediction accuracy in multi-environment scenarios. Overall, the thesis identifies and addresses critical knowledge gaps in deep performance learning, significantly advancing the accuracy of performance prediction.

Rethinking Performance Analysis for Configurable Software Systems: A Case Study from a Fitness Landscape Perspective

Unlocking the Secrets of Software Configuration Landscapes-Ruggedness, Accessibility, Escapability, and Transferability

BestConfig: Tapping the Performance Potential of Systems Via Automatic Configuration Tuning

A Study of Performance Testing in Configurable Software Systems

White-Box Analysis over Machine Learning: Modeling Performance of Configurable Systems

The research of performance evaluation system on manufacturing system with reconfigurable configuration

Deep Configuration Performance Learning: A Systematic Survey and Taxonomy

A survey of fitness landscape analysis for optimization

Identifying Performance-Sensitive Configurations in Software Systems through Code Analysis with LLM Agents

Using Bad Learners to find Good Configurations

White-Box Performance-Influence Models: A Profiling and Learning Approach

Software model refactoring based on performance analysis: better working on software or performance side?

Configuration-space performance anomaly depiction

Transfer learning for performance modeling of configurable systems: An exploratory analysis

Pushing the Boundary: Specialising Deep Configuration Performance Learning

Data-efficient Performance Learning for Configurable Systems

MLOS: An Infrastructure for Automated Software Performance Engineering

Visual analysis of fitness landscapes in architectural design optimization

Performance evaluation modeling for reconfigurable processor

Perf-AL

Rule-based automatic software performance diagnosis and improvement