Experiments · E215c

Is the sampler's temperature error exactly a factor of two?

Yes. Measured on a fine 20 K grid, the factor is 2.000, to within 1.4 %.

In the log: RESULT — the factor of two is 2.000 ± 0.03 (2026-09-22 03:2x)

confirmedDate 2026-09-22 03:2x, as written in the logrung 1 · ordering0 predictions · 1 result paragraphEXPERIMENTS.md lines 15376–15386
exp E215c diagram
What E215c did and how it came out, drawn from this record and the files it names (book/assets/diagrams/exp/E215c.svg).

Results

EXPERIMENTS.md · line 15376

E215c RESULT — the factor of two is 2.000 ± 0.03 (2026-09-22 03:2x)

The exact-Ising eCE (gate RMSE 0.024 meV, J₁ 19.122) swept by ece mcmc on a 20 K grid, 800 → 400 K nominal, 21 states, 6³. Peaks: dE/dT 658.4 K nominal, variance 668.0 — ×2 = 1316.8 / 1335.9 K against the independent L=6 numpy control's 1294–1340 (centre 1317 ± 23) and the thermodynamic 1409.7. Factor to the control: 2.000 (dE/dT), 1.972 (variance); the variance/dE/dT ratio is 2.01, the 2T signature once more. α₁ at the 400 K floor −0.947, E −72.5 meV/atom (B2 −76.5). So PRIMCELL_SITES = 2.0 is measured to 1.4 % on a grid fine enough to resolve it, and the residual 5–7 % between the ×2 numbers and the exact 1409.7 is the L=6 finite-size offset the control carries too, not the sampler. Closes the confirmation the resolution note asked for on 2026-09-21 ("E215c fine grid, factor to a few %").

The full record

EXPERIMENTS.md · lines 15376–15386

E215c RESULT — the factor of two is 2.000 ± 0.03 (2026-09-22 03:2x)

The exact-Ising eCE (gate RMSE 0.024 meV, J₁ 19.122) swept by ece mcmc on a 20 K grid, 800 → 400 K nominal, 21 states, 6³. Peaks: dE/dT 658.4 K nominal, variance 668.0 — ×2 = 1316.8 / 1335.9 K against the independent L=6 numpy control's 1294–1340 (centre 1317 ± 23) and the thermodynamic 1409.7. Factor to the control: 2.000 (dE/dT), 1.972 (variance); the variance/dE/dT ratio is 2.01, the 2T signature once more. α₁ at the 400 K floor −0.947, E −72.5 meV/atom (B2 −76.5). So PRIMCELL_SITES = 2.0 is measured to 1.4 % on a grid fine enough to resolve it, and the residual 5–7 % between the ×2 numbers and the exact 1409.7 is the L=6 finite-size offset the control carries too, not the sampler. Closes the confirmation the resolution note asked for on 2026-09-21 ("E215c fine grid, factor to a few %").

Built with PRISMWebsite and visualizations made using Claude