Activity Reduction and Energy Modeling of Gallager-B LDPC Decoders
Bibliographic record
Abstract
RÉSUMÉ: La réduction de la consommation d’énergie des décodeurs LDPC est une étape importante pour l’adoption de ces codes pour des applications ayant de fortes contraintes énergétiques telles que la protection de mémoires sur puce. Les décodeurs LDPC du type Gallager-B (GaB) sont bien adaptés aux applications nécessitant un haut débit grâce à leur faible complexité. Nous proposons un algorithme de décodage fonctionnellement identique à GaB qui dispose d’une probabilité de transition des messages réduite que nous nommons LA-GaB. Nous proposons aussi une technique dérivée de l’évolution de densité permettant d’estimer ses probabilités de transition de messages. Nous montrons une implémentation ASIC de LAGaB dans une configuration déroulée qui réduit l’énergie dynamique de 50% tout en réduisant la surface de fabrication par rapport à l’implémentation équivalente de GaB standard, sans aucun impact sur la performance de décodage. Nous formulons aussi un modèle d’énergie de haut niveau pour cette implémentation déroulée basé sur les probabilités de transition des messages permettant d’explorer efficacement l’impact des divers paramètres de conception sur l’énergie. ABSTRACT: Reducing the energy consumption of low-density parity-check (LDPC) decoders is an essential step towards increasing the relevance of these codes in energy-constrained applications such as protecting on-chip memories. Gallager-B (GaB) LDPC decoders are well suited for high throughput applications due to their low complexity. We propose the Low-Activity Gallager-B (LA-GaB) decoder, a functional equivalent to GaB with reduced message switching probability, and an adapted density evolution technique capable of predicting its message switching probabilities. We demonstrate an ASIC implementation of LA-GaB in an unrolled configuration achieving a reduction of more than 50% of dynamic energy and lower surface area compared to standard GaB, with no change in decoding performance. We also propose a high-level energy model of this unrolled implementation based on our activity model for rapid design exploration in the energy domain.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 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".