Michael Yang

DOI: https://doi.org/10.5281/zenodo.24197308

We examine two recorded experiments relevant to the proposed dense PFT/HKD Inf 8K graphics engine: a structured dense-training contraction and a matched 8K image-warp benchmark. On a 576-parameter polynomial network, five-node weighted evaluation reduced the median time for twenty dense Adam updates by factors of 1.288, 5.719, and 30.905 relative to full sample evaluation. Against a polynomial-moment implementation the factors were 1.385–1.411. Maximum final-parameter discrepancies were below 2.3 × 10−16. Separately, the 8K PFT JPEG path achieved 1.186 times the speed of direct dense rendering but 0.945 times that of optimized separable dense rendering. These are distinct workloads: the training result is not an end-to-end 8K Renderdon speedup. A self-contained Python appendix reproduces every table from the recorded observations.

Polynomial thinning (PFT); HKD Inf; polynomial loss contraction; Lagrange interpolation; dense neural network training; 8K image rendering; numerical accuracy; performance benchmarking; reproducible computation.

The evidence supports efficient, numerically faithful contraction for this structured dense polynomial loss. It also records an 8K image-warp path whose advantage depends on the baseline: faster than direct dense evaluation but slower than optimized separable evaluation. An end-to-end dense PFT/HKD Inf Renderdon 8K speedup, improved visual accuracy, and general neural-generation applicability remain unestablished by these experiments.

[1] Local project benchmark archive. Polynomial contraction audit: a dense toy network, not a video gen erator. November 5, 2026. Results, source and review: session_checks/pft_dense_review_20261105/ results.json; session_checks/pft_dense_review_20261105/evaluate.py; session_checks/pft_dense_ review_20261105/THEOREM_AND_REVIEW.txt. Unpublished primary experiment record.
[2] Local project benchmark archive. Matched 8K PFT/JPEG benchmark. October 27, 2026, 09:48:18 09:49:42 UTC. benchmark_results/pft_8k_matched_20261027/results.json; benchmark_results/pft_ 8k_matched_20261027/benchmark.py; benchmark_results/pft_8k_matched_20261027/REPORT.txt. Unpub lished primary experiment record.

Michael Yang (2026). Dense PFT / HKD Inf for an 8K Graphics Engine: Measured Speed, Numerical Accuracy, and Disclosure Boundaries. International Journal of Computer Techniques, 13(6), 750–755. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.24197308