Bibliographic record
Abstract
On-chip speeds and integration densities have grown exponentially over the past several decades creating a corresponding demand for high-bandwidth, chip-to-chip communication. Compared with integrated circuit technology, the technologies for chip-packaging, printed circuit boards, and connectors improve at a much slower rate. This results in a big and growing gap between the I/O bandwidth needed and the I/O bandwidth available. Off-chip bandwidth has become a bottleneck in developing high-speed systems. At high data rates, high-frequency losses, reflections and crosstalk severely degrade signal integrity and limit the performance of off-chip links. To combat these issues, designers increasingly rely on on-chip signal processing methods. This thesis explores the effectiveness of equalizing filters for high-bandwidth, point-to- point, off-chip buses. In this work, we combine modelling, optimization and prototyping to demonstrate that linear programming provides practical, effective and flexible basis for designing equalization filters that greatly increase the bandwidth of high-speed buses on printed circuit boards. We first show that the common eye-mask measure of signal integrity is a worst-case performance measure that corresponds to the metric. We show how eye masks can be parameterized to provide a flexible framework for specifying signal integrity trade-offs. We use these parameterized masks to formulate the JQO optimal equalization filter synthesis problem, and show that it can be extended to the unified optimization of pre-equalization, near-end crosstalk cancellation and decision-feedback equalization filters. Our methods work with detailed, realistic channel models and allow the designer to specify practical constraints such as the maximum filter output and' bounds on filter coefficients. Our approach formulates equalization filter synthesis as a linear programming problem. While this makes our approach very flexible, the linear programs that we create can be quite large. To make our methods practical, we implemented a novel linear system solver for use in Mehrotra's interior point linear programming algorithm. Our solver exploits the specific sparsity properties of our optimization problems. We analyze the time and memory requirements of this new implementation as well as its numerical stability.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| 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 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".