mirror of
https://github.com/allaunthefox/Research-Stack.git
synced 2026-08-13 22:00:34 +00:00
Core FixedPoint.lean is now fully Float-free: - Q16_16.sqrt: integer Newton method (intSqrt, 64-step fuel) - Q16_16.log2: bit-position extraction + linear interpolation - Q16_16.expNeg: 7-segment piecewise-linear (~0.02 max error) - Q0_16.log2: bit extraction with 1/ln(2) scaling Removed from core: ofFloat, toFloat, ln (Q16_16); toFloat, ofFloat (Q0_16); q0_64ScaleFloat, ofFloat, toFloat (Q0_64). Created Semantics.FixedPointBoundary for I/O boundary Float conversions. 14 downstream files updated to import boundary module. Build: 3573 jobs, 0 errors (lake build) Co-Authored-By: Allaun Silverfox <bigdataiscoming+9i37y6j2@protonmail.com> |
||
|---|---|---|
| .. | ||
| AdaptiveBlock.lean | ||
| CellCore.lean | ||
| CodingCost.lean | ||
| CompileToPatch.lean | ||
| CompressionPattern.lean | ||
| Core.lean | ||
| HutterContext.lean | ||
| HutterUncompressed.lean | ||
| Metatyping.lean | ||
| PhiRedundancy.lean | ||
| ProofReplay.lean | ||
| QuantumEraserCache.lean | ||
| SignalPolicy.lean | ||
| UnifiedCompression.lean | ||