Regex solving using GPGPU with CUDA
Bibliographic record
Abstract
Regular expression (RE) matching is a computationally intensive task that can benefit from modern, high-performance and concurrent computing.There have already been related optimization efforts, such as HyperScan [12], which is based on SIMD instructions for CPUs, and algorithms like iNFAnt [7] and ASyncAP [23] that target GPUs, improving performance by exploiting the mapping between REs and their finite state machine representations.GPU-based RE acceleration methods, however, can suffer from expensive execution costs when an RE has many initial potential state transitions, and performance heavily depends on ensuring algorithm parameters properly match GPU capabilities.In this thesis, we present a novel study that aims to boost performance and broaden applicability on the GPU side.We introduce a pre-filtering technique that checks the match of simpler RE parts before proceeding to more complex ones.We also optimize the GPU parameters, such as thread occupancy, to avoid naive implementation pitfalls and implement additional optimizations to the state-of-the-art GPU-based algorithm to avoid performance issues caused by edge cases.Our design achieves impressive performance improvement, about 40x faster than iNFAnt and up to 1900x faster than ASyncAP in edge cases, while still maintaining competitive performance in more common cases.The use of our GPU-based optimizations greatly improves the potential for more efficient and versatile RE matching on modern GPUs.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.021 | 0.006 |
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".