Robert Allan JamesandClaude Sonnet 5
cf5b08fb65
every factor interaction, and a raw-data appendix Expanded the campaign-mechanism validation report from a condensed 6-page summary into the full depth Captain Bob asked for: analyze all 9 cells as a conglomerate Latin square, then dive into each cell's own data, then cover every within-ISA and cross-factor interaction explicitly rather than averaging it away. Report structure (127 pages, compiled clean, no undefined references): - Front matter: context, methodology, the SWAP-MTX bug narrative (console-interleaving fix + the Fisher-Yates correctness bug and its fix, both already committed separately) - Layer 1: aggregate 3x3 Latin square (heatmap, invariant-metrics table) - Per-Cell Deep Dive (9 sections): cfg-level distribution, summary table, and a rep-order execution-trajectory chart per cell -- the trajectory charts are what actually visualize the order-dependence finding rather than just stating it - Per-ISA Deep Dive (3 sections): within-architecture seed comparison (violin plots, Kruskal-Wallis, per-factor main effects) - Factor Interactions (6 sections, every pairwise combination of the 4 L8 binary factors): both infer_dec_q and early_exit interaction plots faceted by architecture, plus the three-way factor x factor x architecture significance test - Per-Factor Response (4 sections): linear response by architecture, with an explicit note that a true quadratic term isn't identifiable from this 2-level factorial design - Appendix: full run_id-ordered raw data, all 4,320 rows across all 9 cells, as the primary-source backing for every statistic above Generated programmatically (analyse_stadium_relaunch_fixed.R for the aggregate layer, generate_stadium_deepdive.R for the per-cell/per-ISA/ interaction/appendix layers) rather than hand-authored, since content at this scale needs to be data-driven to stay honest. Also includes analyse_stadium_relaunch.R, the earlier script built against the pre-fix (buggy-shuffle) dataset -- superseded but kept for the record, matching how the underlying data commits were handled. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
43 lines
1.2 KiB
TeX
43 lines
1.2 KiB
TeX
\subsection{Architecture: aarch64}
|
|
\label{sec:isa_aarch64}
|
|
|
|
All three seeds' cells for aarch64 combined: 1{,}440 rows, zero errors. Seed remains the dominant source of variation within this architecture alone (Kruskal-Wallis $H=50.76$, $p=9.51e-12$) -- identical in shape to the other two architectures (see \S\ref{sec:isa_amd64}--\ref{sec:isa_riscv64}).
|
|
|
|
\begin{figure}[h]
|
|
\centering
|
|
\includegraphics[width=0.75\textwidth]{isa_aarch64_seed_violin_light.pdf}
|
|
\caption{aarch64: infer\_dec\_q distribution by seed (violin + inner boxplot).}
|
|
\end{figure}
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\caption{Per-seed summary, aarch64.}
|
|
\begin{tabular}{lrrrr}
|
|
\toprule
|
|
seed & mean & sd & median & early-exit rate \\
|
|
\midrule
|
|
12345 & 56.85 & 29.30 & 68.0 & 0.988 \\
|
|
67890 & 28.09 & 38.34 & 0.0 & 0.990 \\
|
|
13579 & 44.50 & 35.48 & 68.0 & 0.992 \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\end{table}
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\caption{Per-factor main-effect means, aarch64 (all seeds pooled).}
|
|
\begin{tabular}{lrrrr}
|
|
\toprule
|
|
factor & mean (hi) & mean (lo) & difference \\
|
|
\midrule
|
|
entropy & 43.80 & 42.50 & 1.30 \\
|
|
coeff.\ of variation & 43.11 & 43.18 & -0.07 \\
|
|
temporal decay & 43.39 & 42.90 & 0.49 \\
|
|
stability & 43.52 & 42.77 & 0.75 \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\end{table}
|
|
|
|
\clearpage
|
|
|