Sound Static Data Race Verification for C: Is the Race Lost?
Bibliographic record
Abstract
This artifact contains the benchmarks, tools and scripts for reproduction, along with our reference results used for the paper. Contents The reproduction package contains materials for reproducing Tables 2, 4, 5, 10, and 14 from the paper. These tables provide the data supporting research questions 2 and 3, as well as additional evaluation results. We provide two versions of the artifact: The source version includes benchmarks, scripts and reference results such that they can easily be accessed and reused outside of the virtual machine. The virtual machine version additionally includes tools and their dependencies such that the results can be reproduced by execution. The source version contains: README.md/README.pdf — This file. concrat-benchmarks/ — Concrat benchmarks (RQ 3) and execution scripts. results-paper/ — Reference results used for Table 2. extracted-micro-benchmarks/ — Extracted micro-benchmarks (with their racy variations) and execution scripts (RQ 2). results-paper/ — Reference results used for Table 4 (Finding 2). concrat-benchmarks-excluded/ — Excluded Concrat benchmarks (RQ 3). sv-benchmarks/ — SV-COMP 2023 NoDataRace-Main category benchmarks. joern/ — Joern scripts for Table 5 (RQ 3). concrat-benchmarks-paper/ — Reference results for Concrat benchmarks used for Table 5 (Finding 3). concrat-benchmarks-excluded-paper/ — Reference results for excluded Concrat benchmarks used for Table 5 (Finding 3). sv-benchmarks-paper/ — Reference results for SV-COMP benchmarks used for Table 5 (Finding 3). extracted-micro-benchmarks-paper/ — Reference results for extracted micro-benchmarks used for Table 14. sv-benchmarks.sh — Script to download SV-COMP 2023 NoDataRace-Main category benchmarks. tools/download.sh — Script to download SV-COMP 2023 tools from their reproduction packages. properties/no-data-race.prp — Property file for executing SV-COMP tools. tsan-races/ — Scripts to run ThreadSanitizer on Concrat benchmarks. logs/ — Reference results used for Table 2 and Table 10. The virtual machine version contains all of the above in /home/vagrant, but also: concrat-benchmarks/ results-test/ — Results from kick-the-tires (initially empty). results/ — Full evaluation results (initially empty). results-reduced/ — Reduced evaluation results (initially empty). extracted-micro-benchmarks/ results-test/ — Results from kick-the-tires (initially empty). results/ — Full evaluation results (initially empty) (Finding 2). results-reduced/ — Reduced evaluation results (initially empty) (Finding 2). joern/ concrat-benchmarks/ — Results for Concrat benchmarks (initially empty) (Finding 3). concrat-benchmarks-excluded/ — Results for excluded Concrat benchmarks (initially empty) (Finding 3). sv-benchmarks/ — Results for SV-COMP benchmarks (initially empty) (Finding 3). tools/ (subdirectories) — Downloaded SV-COMP 2023 tools from their reproduction packages. Hardware Dependencies The executable artifact is a VirtualBox virtual machine, because BenchExec does not run in Docker. Full evaluation requires: 8 CPU cores, 26 GB RAM, 7 GB disk space, ~2 days and 15 hours. Considering the significant runtime, we also provide a reduced evaluation. Reduced evaluation requires: 8 CPU cores, 16 GB RAM, 7 GB disk space, ~2 hours.
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 | 0.011 | 0.085 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.006 | 0.015 |
| Open science | 0.005 | 0.005 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.054 | 0.026 |
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".