Bibliographic record
Abstract
Performance makes or breaks a software system. Severe performance issues render systems unusable; even small performance degradation can be costly. Google finds that a 0.5-second delay in page load time has caused a 20% drop in repeat traffic [74]. Despite its importance, there are many obstacles to building performant systems. Modern software systems have complex performance characteristics that are difficult to understand, and users and developers still invest significant time manually diagnosing critical performance issues. The goal of my dissertation work is to help developers build performant software systems, by understanding the performance characteristics of the software systems comprehensively, and creating automated techniques to improve performance. This dissertation consists of two pieces of work, the first piece of work is a study that discovered that many of Linux’s core kernel operations had slowed down over time or had unstable performance due to new features and security guarantees being added, as well as misconfigurations. Despite unavoidable performance trade-offs, we found much of the slowdown can be improved through proactive testing, diagnosis, and optimization or avoided through custom kernel configuration. Motivated by the previous study, the second piece of work creates an automated technique—relational debugging—to diagnose the root causes of performance issues. Relational debugging is inspired by the observation that existing root cause diagnosis tools make fundamental assumptions that are broken by performance issues. Such tools only work for functional bugs that cause clear-cut failures, like crashes, when the expected performance outcome must be determined relative to input workloads. Moreover, existing tools only capture absolute root causes, e.g., broken invariants or predicates, whereas performance root causes tend to be abnormal relative occurrences of events. Based on these insights, we create the relational debugging algorithm, which caters to the “relative” characteristics of performance issues, and we implemented the algorithm with Perspect [105], which can effectively diagnose complex performance issues that human developers struggle with and existing tools are ineffective against.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.016 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".