A Customizable Lightweight STM for Irregular Algorithms on GPU
Bibliographic record
Abstract
Irregular algorithms are often encountered in highly data-centric application domains. These algorithms operate on irregular data structures such as sparse graphs with irregular access patterns, which may also modify the underlying topology unpredictably. High computational time and inherent data parallelism present in these algorithms motivate the use of GPUs for speeding things up, however there are challenges for their efficient implementations due to: difficulty in protecting the shared data consistency in the presence of concurrent dynamic transactions; irregular access patterns due to unstructured data structures; and dynamic structural modifications of the underlying topology. One approach to overcome these challenges is to use Software Transactional Memory (STM). However, overly complex design and implementations of contemporary STM-based approaches and lack of proper framework to employ them in conjunction with the irregular algorithms stalls their adoption by the programming community. To overcome some of these challenges, this research proposes a lightweight STM with a simple design (Lite GSTM), based on a lock stealing algorithm, and an associated extensible framework to hide the complexity of the STM from a programmer. The framework is extensible by allowing plug-ins of customized STMs designed for different needs of transactions. The use of the framework is elaborated with two use cases which employ completely different irregular algorithms, however, have some common features: the underlying data structure is a graph, and the graph is structurally modified (coarsened) unpredictably in the course of execution. The paper presents the performance comparisons of the STM-based implementations with respect to their sequential and non-STM based counterparts, which show promising results.
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 | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
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".