StarForth / StarshipOS – Full DOE Report

Generated: 2025-11-26 10:50:41.480068

Runs: 38400 | Configurations: 128 | Replicates: 30

1. ns_per_word Distribution

2. ANOVA – Main Effects

Analysis of Variance Table

Response: ns_per_word
                         Df     Sum Sq    Mean Sq    F value    Pr(>F)    
L1_heat_tracking          1 7.4240e+16 7.4240e+16  1152.4356 < 2.2e-16 ***
L2_rolling_window         1 4.3707e+14 4.3707e+14     6.7847  0.009198 ** 
L3_linear_decay           1 1.3341e+15 1.3341e+15    20.7091 5.363e-06 ***
L4_pipelining_metrics     1 2.9975e+18 2.9975e+18 46530.2403 < 2.2e-16 ***
L5_window_inference       1 1.5384e+14 1.5384e+14     2.3880  0.122276    
L6_decay_inference        1 4.7490e+13 4.7490e+13     0.7372  0.390564    
L7_adaptive_heartrate     1 5.9415e+13 5.9415e+13     0.9223  0.336875    
replicate               299 1.6852e+16 5.6362e+13     0.8749  0.941246    
Residuals             38093 2.4539e+18 6.4420e+13                         
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

3. ANOVA – 2-Way Interactions

Analysis of Variance Table

Response: ns_per_word
                                               Df     Sum Sq    Mean Sq
L1_heat_tracking                                1 7.4240e+16 7.4240e+16
L2_rolling_window                               1 4.3707e+14 4.3707e+14
L3_linear_decay                                 1 1.3341e+15 1.3341e+15
L4_pipelining_metrics                           1 2.9975e+18 2.9975e+18
L5_window_inference                             1 1.5384e+14 1.5384e+14
L6_decay_inference                              1 4.7490e+13 4.7490e+13
L7_adaptive_heartrate                           1 5.9415e+13 5.9415e+13
replicate                                     299 1.6852e+16 5.6362e+13
L1_heat_tracking:L2_rolling_window              1 3.1983e+14 3.1983e+14
L1_heat_tracking:L3_linear_decay                1 2.5183e+15 2.5183e+15
L1_heat_tracking:L4_pipelining_metrics          1 1.0587e+15 1.0587e+15
L1_heat_tracking:L5_window_inference            1 5.1359e+13 5.1359e+13
L1_heat_tracking:L6_decay_inference             1 3.3666e+13 3.3666e+13
L1_heat_tracking:L7_adaptive_heartrate          1 2.0141e+14 2.0141e+14
L2_rolling_window:L3_linear_decay               1 1.0882e+15 1.0882e+15
L2_rolling_window:L4_pipelining_metrics         1 5.2174e+14 5.2174e+14
L2_rolling_window:L5_window_inference           1 1.5498e+15 1.5498e+15
L2_rolling_window:L6_decay_inference            1 3.7362e+12 3.7362e+12
L2_rolling_window:L7_adaptive_heartrate         1 1.9174e+14 1.9174e+14
L3_linear_decay:L4_pipelining_metrics           1 6.5727e+14 6.5727e+14
L3_linear_decay:L5_window_inference             1 2.7786e+12 2.7786e+12
L3_linear_decay:L6_decay_inference              1 7.1187e+10 7.1187e+10
L3_linear_decay:L7_adaptive_heartrate           1 1.9866e+13 1.9866e+13
L4_pipelining_metrics:L5_window_inference       1 4.8622e+14 4.8622e+14
L4_pipelining_metrics:L6_decay_inference        1 6.9978e+13 6.9978e+13
L4_pipelining_metrics:L7_adaptive_heartrate     1 1.7744e+13 1.7744e+13
L5_window_inference:L6_decay_inference          1 2.5715e+15 2.5715e+15
L5_window_inference:L7_adaptive_heartrate       1 5.5548e+15 5.5548e+15
L6_decay_inference:L7_adaptive_heartrate        1 2.0477e+14 2.0477e+14
Residuals                                   38072 2.4368e+18 6.4006e+13
                                               F value    Pr(>F)    
L1_heat_tracking                             1159.8939 < 2.2e-16 ***
L2_rolling_window                               6.8286  0.008974 ** 
L3_linear_decay                                20.8431 5.000e-06 ***
L4_pipelining_metrics                       46831.3723 < 2.2e-16 ***
L5_window_inference                             2.4035  0.121074    
L6_decay_inference                              0.7420  0.389035    
L7_adaptive_heartrate                           0.9283  0.335316    
replicate                                       0.8806  0.931891    
L1_heat_tracking:L2_rolling_window              4.9968  0.025399 *  
L1_heat_tracking:L3_linear_decay               39.3452 3.589e-10 ***
L1_heat_tracking:L4_pipelining_metrics         16.5406 4.772e-05 ***
L1_heat_tracking:L5_window_inference            0.8024  0.370377    
L1_heat_tracking:L6_decay_inference             0.5260  0.468302    
L1_heat_tracking:L7_adaptive_heartrate          3.1468  0.076085 .  
L2_rolling_window:L3_linear_decay              17.0012 3.743e-05 ***
L2_rolling_window:L4_pipelining_metrics         8.1514  0.004305 ** 
L2_rolling_window:L5_window_inference          24.2128 8.662e-07 ***
L2_rolling_window:L6_decay_inference            0.0584  0.809088    
L2_rolling_window:L7_adaptive_heartrate         2.9957  0.083492 .  
L3_linear_decay:L4_pipelining_metrics          10.2689  0.001354 ** 
L3_linear_decay:L5_window_inference             0.0434  0.834953    
L3_linear_decay:L6_decay_inference              0.0011  0.973396    
L3_linear_decay:L7_adaptive_heartrate           0.3104  0.577449    
L4_pipelining_metrics:L5_window_inference       7.5965  0.005851 ** 
L4_pipelining_metrics:L6_decay_inference        1.0933  0.295747    
L4_pipelining_metrics:L7_adaptive_heartrate     0.2772  0.598533    
L5_window_inference:L6_decay_inference         40.1755 2.347e-10 ***
L5_window_inference:L7_adaptive_heartrate      86.7853 < 2.2e-16 ***
L6_decay_inference:L7_adaptive_heartrate        3.1993  0.073680 .  
Residuals                                                           
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

4. All 128 Configurations

4a. Ranked Bar Chart

4b. 16x8 Performance Heatmap (L1234 vs L567)

4c. Variance Heatmap (CV%)

Full data: all_128_configurations.csv

5. OPTIMAL STATE ANALYSIS (Speed + Stability)

The optimal performer is NOT just the fastest - it's the configuration that achieves both fast execution AND steady-state behavior, representing the true optimal dynamic system equilibrium.

5a. Pareto Frontier: Speed vs Stability

Lower-left corner = Pareto-optimal (fast AND stable). Top 5% configs circled in red.

5b. Optimality Score Distribution

5c. Bell Curve Fit with Top 5% Highlighted

Green shaded region shows the top 5% of configurations by optimality score.

5d. Top 5% Configuration Breakdown

Component scores for each top 5% config: Speed, Stability, and Overall.

5e. Loop Enrichment in Top 5%

Which feedback loops are over/under-represented in optimal configurations?

5f. Top 5% Optimal Configurations

# A tibble: 7 × 5
  config_bits   mean_ns cv_pct optimality_score optimality_percentile
  <chr>           <dbl>  <dbl>            <dbl>                 <dbl>
1 0110001     32293075.   14.4           0.0535                 100  
2 0010110     31962118.   14.6           0.0654                  99.2
3 0010111     31673508.   14.8           0.0808                  98.4
4 0100110     32002506.   15.1           0.102                   97.7
5 0100011     31592404.   15.1           0.103                   96.9
6 0010001     31921582.   15.3           0.118                   96.1
7 1010001     34423368.   15.0           0.139                   95.3

Full ranking: optimal_configurations_ranked.csv

5g. Top 5% vs Baseline (0000000) Comparison

Baseline config (0000000): mean = 3.17e+07 ns/word, CV = 17.34%

Speedup with Statistical Significance (T-test)

Dual Improvement: Speed + Stability

Detailed Comparison Table

      config_bits  mean_ns speedup_vs_baseline   cv_pct cv_improvement
t...1     0110001 32293075        -1.960683620 14.39137       2.945347
t...2     0010110 31962118        -0.915735239 14.61180       2.724917
t...3     0010111 31673508        -0.004490247 14.83188       2.504837
t...4     0100110 32002506        -1.043253850 15.10276       2.233964
t...5     0100011 31592404         0.251583952 15.12602       2.210701
t...6     0010001 31921582        -0.787748632 15.30973       2.026988
t...7     1010001 34423368        -8.686774250 15.01970       2.317025
          t_pvalue t_significant
t...1 1.354071e-01         FALSE
t...2 4.861458e-01         FALSE
t...3 9.972814e-01         FALSE
t...4 4.343161e-01         FALSE
t...5 8.496873e-01         FALSE
t...6 5.568347e-01         FALSE
t...7 5.175982e-10          TRUE

Full comparison: top5pct_vs_baseline.csv

6. All 21 Pairwise Interaction Plots

1. L1 x L2

2. L1 x L3

3. L1 x L4

4. L1 x L5

5. L1 x L6

6. L1 x L7

7. L2 x L3

8. L2 x L4

9. L2 x L5

10. L2 x L6

11. L2 x L7

12. L3 x L4

13. L3 x L5

14. L3 x L6

15. L3 x L7

16. L4 x L5

17. L4 x L6

18. L4 x L7

19. L5 x L6

20. L5 x L7

21. L6 x L7

ANOVA results: interaction_anova_results.csv

6. ns_per_word by Individual Factors

L1_heat_tracking

L2_rolling_window

L3_linear_decay

L4_pipelining_metrics

L5_window_inference

L6_decay_inference

L7_adaptive_heartrate

7. Metric Correlation Heatmap

8. Factor → Metric Effect Map

9. Key Metrics vs Factors