Rounded parameters for a near-halving ordinary extractor #
At target output length b, use source width 8*b, subtract three ceiling
logarithms and fifteen from its binary logarithm to choose the depth, and
round the reserve upward to obtain at least b output bits. The size guard
makes the depth subtraction exact. These are total natural-number definitions;
their useful finite inequalities and eventual validity are separate results.
Ceiling logarithm controlling the selected seed budget.
Equations
- Algebraic.Cutwidth.Extractor.gammaBlockLog b = Nat.clog 2 (b + 1)
Instances For
Leave a cubic logarithmic output quantum below the target width.
Equations
- Algebraic.Cutwidth.Extractor.gammaBlockDepth b = Nat.log 2 b - 3 * Nat.clog 2 (Algebraic.Cutwidth.Extractor.gammaBlockLog b + 1) - 15
Instances For
The local error exponent paying for final error one quarter.
Equations
Instances For
Number of output bits contributed by one unit of the leaf reserve.
Equations
Instances For
Round the leaf reserve upward so the complete output covers b bits.
Equations
Instances For
The common final output length of every leaf extractor.
Equations
Instances For
The actual input entropy threshold in bits for the rounded schedule.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The logarithmic depth subtraction does not truncate at zero.
Equations
- Algebraic.Cutwidth.Extractor.GammaBlockSizeGuard b = (3 * Nat.clog 2 (Algebraic.Cutwidth.Extractor.gammaBlockLog b + 1) + 15 ≤ Nat.log 2 b)