3.051 mV syntheticNominal matched read sigmaUNKNOWNInspect metric contract
300 / 300 defaultCommon cellsUNKNOWNInspect metric contract
<= 1e-9 mV^2Variance closureUNKNOWNInspect metric contract
SIMULATION / SA-MEM-SRAM-VARIATION-COVERAGE-060
SRAM variation and corner coverage
SRAM Variation Coverage
Separate what moved from what was missing. Inspect fixed condition, lot, wafer, spatial, local-mismatch, noise, and coverage effects before interpreting an SRAM margin proxy.
EXPLORE
Start with the observable.
Use this page when the simulation is the primary artifact. The linked lesson explains the concept; the linked lab keeps the visual replay surface available.
- Hold the fixed condition and switch among lot-, wafer-, spatial-, local-, and noise-dominant presets.
- Compare pooled-total, common-set, all-valid, and condition-matched views while watching every denominator.
- Inspect PD/PG/PU direction coefficients, covariance remainder, and why a fixed corner is not automatically a sigma.
SYNTHETIC HIERARCHY LAB / V128
Separate what moved from what was missing.
A known-ground-truth condition-lot-wafer-site hierarchy keeps variance, coverage, and uncertainty denominators in the same ledger.
Current tab onlyillustrative-original-surrogateNo PDK, production row, upload, storage, or network writeFixed condition is not sigma. This lab does not predict yield.Covariance remainder closes total variance; Evidence stop stays explicit.
Selected sigma3.051 mV
Population sigma3.048 mV
Common cells300 / 300
Closure error0.0e+0
Variance divisor
Condition matchedValid replicate rows at one condition, restricted to cells present in every condition.
SELECTED DISTRIBUTION
Read margin proxy
Mean120.000 mV
Minimum110.202 mV
P05114.900 mV
Median120.054 mV
P95125.097 mV
Maximum129.111 mV
Degrees of freedom599
First-order sigma relative uncertainty2.89%
iid normal reference only; hierarchy and repeated observations reduce effective independence
Selected values600
Selected cells300
Selected cell-condition pairs300
Valid / eligible observations2,400 / 2,400
RUN CARD
Conditions before conclusions.
The browser generates a known-truth synthetic hierarchy in the current tab. It does not load production rows, expose a PDK, map fixed corners to sigma, predict yield, or provide signoff.
Primary knobs
- Seed and hierarchy source sigmas
- measurement-noise sigma
- cell-condition and replicate missingness
- metric, condition, analysis set, and divisor
Saved outputs
- 18-bin selected distribution
- six-component variance ledger plus covariance remainder
- condition and coverage matrices
- PD/PG/PU balance views
- deterministic JSON and CSV
EXPORT
Download the public artifacts.
These exports are the supported public files for review. PDK files and restricted third-party material are not redistributed.
V128 public model contract
/data/semiagora-sram-variation-corner-coverage-v1.json
Open file ->METHODS
How the number was made.
Methods are written for citation discipline: the extraction rule matters as much as the plotted value.
Condition is a fixed synthetic direction; wafer is nested in lot, site in wafer, and replicate in cell-condition.
Keep this method attached when reusing the figure or metric.
Total variance closes through six diagonal component variances plus every pairwise covariance term.
Keep this method attached when reusing the figure or metric.
Common-set and all-valid views use distinct, always-visible denominators; missing values never become zero.
Keep this method attached when reusing the figure or metric.
LIMITS
What this page does not claim.
The MVP is useful because the limits are visible. These statements prevent a teaching simulation from being cited as silicon evidence.
Every row, condition, component, strength, margin value, missingness flag, and distribution is deterministic synthetic teaching data.
Carry this boundary into any derivative note, paper draft, or slide.
The balanced hierarchy and orthogonalized fields are designed for interpretability; they are not a foundry sampling plan or covariance model.
Carry this boundary into any derivative note, paper draft, or slide.
The four fixed conditions are deliberate directions, not sampled tails, PDK corners, or sigma equivalents.
Carry this boundary into any derivative note, paper draft, or slide.
PD, PG, and PU strengths and the read, write, and inverter-skew values are linear original proxies with arbitrary visible coefficients and no calibration.
Carry this boundary into any derivative note, paper draft, or slide.
The first-order sigma uncertainty reference assumes iid normal observations; nested and repeated rows are not independently sampled.
Carry this boundary into any derivative note, paper draft, or slide.
Finite subsets and missingness can create component covariance, so diagonal component variances alone need not equal total variance.
Carry this boundary into any derivative note, paper draft, or slide.
The lab cannot establish physical causality, statistical representativeness, measurement adequacy, process capability, Vmin, failure probability, array yield, recipe transfer, qualification, signoff, or production readiness.
Carry this boundary into any derivative note, paper draft, or slide.
Researcher-facing businesses remain deferred to a separately verified V200 contract.
Carry this boundary into any derivative note, paper draft, or slide.
EVIDENCE
Trace the claim back to artifacts.
The evidence route remains the catalog-level browser; this page is the simulation-level reading card.
Every row retains its exact generated contribution ledger and deterministic digest.
SA-MEM-SRAM-VARIATION-COVERAGE-060
The sigma uncertainty reference is iid-normal only; hierarchical rows are not all independent.
SA-MEM-SRAM-VARIATION-COVERAGE-060
Five measured-replacement slots state what is required before a physical, yield, or signoff claim.
SA-MEM-SRAM-VARIATION-COVERAGE-060
Source spine
These links anchor the public teaching model. They do not convert the replay into a qualified process recipe or signoff result.
How can I account for nested variation?hierarchy and denominator vocabularyWhat are variance components?variance-component and uncertainty boundaryConfidence interval approach for standard deviationsigma-estimate uncertainty contextNIST TN 1297: Reporting Uncertaintyreporting discipline onlyMatching Properties of MOS Transistorscategory-level semiconductor variability context onlyAnalysis and Decomposition of Spatial Variation in Integrated Circuit Processes and Devicescategory-level semiconductor variability context onlyStatistical Analysis of SRAM Cell Stabilitycategory-level semiconductor variability context onlyLER/LWR to SRAM variability chaininternal public continuity
Public execution boundary
No server-side live ngspice execution, uploads, accounts, payments, comments, newsletter signup, or job/event submissions are active on this page.