Commit graph

2,076 commits

Author SHA1 Message Date
Meta Vens
15730d4e99 cultmath: span tests: the float arm of the near-zero midpoint test carries the accumulation bound of its octave reads; previous commit's new test failed on 7e-9 of float noise (builds, one new test red)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-15 23:13:30 +02:00
Meta Vens
8939af134f cultmath: span tests: long path from rest, near-zero midpoint under a point span and a loose speed bound, fold's float arm read from span_ridged's own field
The fold test's float arm no longer re-spells the octave coordinate (OneOctave): each octave's float value is the difference of span_fbm / span_ridged at o + 1 and o octaves, with a stated bound on what the accumulating float sum loses.
New fixed span at halfLength 6 pins the Gl reach operator (nu / halfLength fails); a near-zero midpoint on a stationary path (halfLength 0, and halfLength 1 under a loose speed bound) pins the halfLength-0 fold rule and S_o's reach term.
Builds: yes (to be confirmed by the Yggdrasil job).

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-15 23:12:41 +02:00
Meta Vens
3de67b9f9a cultmath: span tests: octave-sum identity, stretched-velocity and transformed-acceleration folds, zero axis with stretch below 1, bands lifted through octaveStep
Kills the Stryker survivors of the first allowance commit: the velocity stretch sign, the norm at
a zero axis, gain-weight and acceleration-factor arithmetic, the octaveStep term of S_o. Not yet
built or run.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 21:25:22 +02:00
Meta Vens
ffb25b6657 cultmath: span tests: MaxMagnitude as a loop (a lambda cannot capture a struct's this); previous commit did not build the test project
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 20:51:58 +02:00
Meta Vens
52d25740ac caching: MessagePack and JSON shapes for SpanPath and SpanStage
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 20:49:31 +02:00
Meta Vens
0a1b49a231 cultmath: span stages take SpanPath and SpanStage and widen curvature and fold by float rounding allowances
span_fbm/span_ridged/span_stage lose the 14-argument signatures for SpanPath, SpanStage,
octaves, lacunarity, gain. Per octave S_o and e_n, e_g, e_h (af_snoise's derivation plus the
Hessian's 2^-10) widen the local Hessian tier, the gradient bound and the fold travel. Tests
compare against the exact double kernel in coordinate bands, add the pinned rank-one span,
widen the stage parameter ranges, and state the lab-form mapping. Not built or run yet.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 20:49:30 +02:00
Meta Vens
f722f7c177 cultmath: span fold test for the acceleration term alone
SpanFoldCoversAccelerationAlone draws spans with zero midpoint velocity and an acceleration along the gradient
that carries n across zero, so the fold's acceleration term and its gradient bound are the only thing holding
the branch departure under fold; two hand mutants survived without it.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 21:03:52 +02:00
Meta Vens
6a3132f139 cultmath: span stage value and gradient against an independent octave assembly
SpanStageGradientMatchesDifferences now also compares field.w and field.xyz with an octave sum assembled
in the test from Stage.Point and snoise_grad, so shift, offset, octaveStep, M, lacunarity and gain are pinned
numerically, which a value/gradient self-consistency check cannot do.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:53:37 +02:00
Meta Vens
427d81feda GameCult.Caching: SpanNoise MessagePack and JSON shapes
MessagePack [field, curvature, fold] beside CultPhasor in CultMathResolver, JSON {field, curvature,
fold} in CultMathJson; round-trip, hex and JSON golden cases beside CultPhasor's.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:48:44 +02:00
Meta Vens
c2722c651b cultmath: span_fbm and span_ridged return SpanNoise (field, curvature, fold)
The octave stages of a warp chain over a span, in C#: stretched, shifted, offset octave sums whose
field equals fbm_grad and ridged_grad bit for bit at the identity transform and whose curvature
and fold bounds derive from SNOISE_LIPSCHITZ, SNOISE_HESSIAN and SNOISE_THIRD only (local tier from
the midpoint Hessian's Frobenius norm plus the third-derivative term, capped by the global pin).
SpanStageTests encloses the Hessian norm and the branch departure over seeded spans in both regimes
and checks the gradient against differences and the stretch/axis conventions. README and design.md
state the span, the transform and the bound rules. No shader twin: stage-shaders is part 2.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:48:43 +02:00
Meta Vens
6fd60a979f cultmath: integral_clamp_linear normalises lo before the flat branch (gpu-clamp-lo-above-hi-undefined)
lo = min(lo, hi) now precedes the b == 0 branch in C# and HLSL (Unity copy byte-identical, GLSL
regenerated), so no clamp ever sees lo above hi; GLSL leaves that undefined. The b == 0 path used to
call clamp(a, lo, hi) with the raw bounds. New tests: the flat and sloped lo > hi values, the
statement-order check on the committed HLSL and GLSL, and an appended glsl-parity.json family of lo > hi
and infinite-bound edge cases (every earlier case byte-identical).

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:42:04 +02:00
Meta Vens
b042bf19be cultmath: integral_clamp_linear is finite for infinite clamp bounds (integral-infinite-bound-nan)
Each held stretch contributes only when its length is positive, so an infinite bound times a zero
length never forms: relu (hi = +inf) and its mirrors give the finite integral instead of NaN, in C#,
HLSL (Unity copy byte-identical) and the generated GLSL. Finite cases are unchanged.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:39:27 +02:00
Meta Vens
38a2f50eb0 cultmath: closed-form third tensor, SNOISE_THIRD re-pinned to 535.8009 (third-provenance-fd-noise)
ThirdTensor is now snoise_hessian's per-corner derivation differentiated once more, in double, on the
kernel's lattice, tie order and hash; central differences of the float Hessian are gone from the
measurement and survive only as the oracle in ThirdTensorMatchesDifferencesOfTheHessian (Richardson
extrapolated differences of the double Hessian). MeasureThird on the closed form: refined 487.0917,
x 1.10 = 535.8009 (the differenced tensor read 488.1047, its roundoff). SNOISE_THIRD, ThirdWitness,
CULTMATH_SNOISE_THIRD (both HLSL copies), the generated GLSL constant and the design.md provenance
paragraph follow.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:37:36 +02:00
Meta Vens
6e360cf56d cultmath: shader-constant pin reads the mirror's private const fields
The compiled mirror declares its static consts without public access; the first version of
EveryShaderConstantEqualsItsCSharpTwin found none and failed. Reflect NonPublic too.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:37:35 +02:00
Meta Vens
5588e95fa2 cultmath: pin every shader constant to its C# twin (shader-third-constant-unpinned)
EveryShaderConstantEqualsItsCSharpTwin reflects each CULTMATH_* literal of the compiled HLSL
mirror and asserts it is bit-equal to the math constant of the same name, so CULTMATH_SNOISE_THIRD
(and every other constant) edited in one runtime fails.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 20:26:51 +02:00
Meta Vens
2fd9431a76 cultmath-tapes map: target r2 (tape, affine forms, noise pins, stage functions, span framework ground, lattice node); stale parked lines swept
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 20:10:40 +02:00
Meta Vens
1a12bc2d82 cultmath design: SNOISE_HESSIAN re-pinned at 55.642143, snoise_hessian float3x3, SNOISE_THIRD
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 13:36:15 +02:00
Meta Vens
4784c25cc6 Merge hands/cultmath-expr-ir 2026-10-14 13:35:46 +02:00
Meta Vens
9282836fa5 Merge hands/cultmath-snoise-hessian 2026-10-14 13:35:45 +02:00
Meta Vens
d47811c323 ExprGraph resolver: nil is nil, decode refuses non-canonical encodings, pin nested shape, nil-name and depth rules
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 07:54:32 +02:00
Meta Vens
646cc59426 ExprGraph lists are read-only views over private arrays; test short evaluator inputs
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 07:54:31 +02:00
Meta Vens
9bb36a47bf cultmath: regenerate the af_snoise and af_fbm fixture families for the re-pinned SNOISE_HESSIAN
Their results carry the curvature constant; the platform-bound families stay as generated on Windows.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 07:13:49 +02:00
Meta Vens
2c5f71b959 Close the Stryker and hand-mutant survivors: port-name alphabet, refusal messages, zero outputs, column shape, substitution checks, self operand
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 07:01:34 +02:00
Meta Vens
0e26f1e059 cultmath: regenerate CultMath.glsl and append the snoise_hessian and integral_clamp_linear fixture families
CultMath.glsl is GlslLowering's output (CULTMATH_WRITE_GLSL=1, run on Yggdrasil); the fixture keeps every
earlier case byte for byte (the platform-bound families stay as generated on Windows) and appends the three
new families, which depend on no libm.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:55:40 +02:00
Meta Vens
e3ea6d08d4 cultmath hlsl: type-neutral comment on snoise_hessian's matrix layout (lowers to GLSL unchanged in meaning)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:54:48 +02:00
Meta Vens
97de733f1b cultmath: pin SNOISE_THIRD (536.9152) and re-pin SNOISE_HESSIAN (55.642143) on the analytic Hessian
MeasureThird on the tied kernel: refined 488.1047, x 1.10 = 536.9152, witness (-168.3755, -3.8205261,
-109.16948). MeasureHessian with the analytic Hessian: refined 50.58377, x 1.10 = 55.642143, witness
(-13.7920685, -9.672134, 61.563934); the central-difference value 56.050385 carried the quotient's
roundoff. README bullet for the span primitives. CultMath.glsl and the fixture are not yet regenerated.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:53:08 +02:00
Meta Vens
00a477e387 cultmath tests: snoise second-difference check tolerates kernel-edge points by a stated fraction
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:51:42 +02:00
Meta Vens
f7c228a98f Research: lattice-site-weight corrected; the fauna probe weights at the cell centre (the cellc variant), not the jittered site
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:48:38 +02:00
Meta Vens
276484e15b cultmath tests: second differences of snoise compared against a two-level Richardson extrapolation
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:45:41 +02:00
Meta Vens
b8a5a9cc14 cultmath tests: ThirdNorm ascends from the six strongest starts only (cost of MeasureThird)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:43:57 +02:00
Meta Vens
0a4a14dd5e Test that Substitute carries params, constants and picks
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:40:09 +02:00
Meta Vens
38473f3fa5 cultmath: snoise_hessian, integral_clamp_linear and SNOISE_THIRD scaffold (not yet measured or regenerated)
math.Span.cs adds the analytic Hessian of snoise(float3) on snoise_grad's tied kernel and the closed-form
integral of a clamped linear ramp; HLSL twins and the Unity copies follow, HessianNorm now reads the
library Hessian, the GLSL lowering gains float3x3 to mat3, and MeasureThird/ThirdConstantPinsSampled and
SpanPrimitiveTests land. SNOISE_THIRD is a placeholder (1) until MeasureThird runs; CultMath.glsl and the
fixture are not yet regenerated. Does not pass its tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:31:43 +02:00
Meta Vens
59cf29032e Research: span-chain shape (Taylor forms, Mitchell, Knoll, MPR, Engel) and site-ground field (WebGL2 splats, OIT, 3DGRT, EVER)
Eyes pass (2026-10-14) serving cultmath-tapes❓span-chain-shape and
cultmath-tapes❓site-ground-field.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:27:08 +02:00
Meta Vens
21281a39d6 Research: where a weighted lattice cell gets its weight (literature, CultMath today, 2D probe)
Eyes pass (2026-10-14) serving cultmath-tapes❓lattice-site-weight-source.
Key measurement: a query-point weight is common to all candidates and never
changes the winner; site and cell-centre weights differ from each other on
11-15% of pixels; no variant pops in space or time while the window holds
the winner; the window must scale with the weight ratio.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:24:47 +02:00
Meta Vens
9545c80c56 cultmath-tapes map: Pass 8 (volumetric prioritized: frontier, Lipschitz history, Span Lab pieces and owners, curl note, cut sequence)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:20:36 +02:00
Meta Vens
ed2c5843e4 Tests for CultMath.Expr: evaluator equals direct composition, canonical builds, substitution, validator rules, resolver round trip
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:09:31 +02:00
Meta Vens
dfc720e754 Research: tape register model prior art (Fidget, MPR, GPU indexed locals) and body-Voronoi metrics
Eyes pass (2026-10-14) serving cultmath-tapes❓tape-register-model,
with a scope addition on weighted diagrams over gravity bodies and metric
parameterisation for question lattice-site-weight-source.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:05:27 +02:00
Meta Vens
defe2b80de Document CultMath.Expr: design.md Expressions section, resolver README
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:05:19 +02:00
Meta Vens
dc302a96a7 CultMath.Expr: scalar SSA expression graph, builder with substitution, one validator, evaluator, MessagePack shape
ExprGraph (immutable class, one validator in Create), ExprBuilder (hash-consing, dead-node removal, Substitute), ExprOps (the one op table), ExprEvaluator, ExprBindings, ExprGraphException; CultMathExprResolver appended to the default chain after CultMathResolver.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 06:05:18 +02:00
Meta Vens
b136ad9bd4 Merge hands/cultmath-cycles-converters: Cycles shader-node semantics in CultMath, exact to unfused Cycles
Cut cultmath-tapes cycles-converters (spec r2, attempts h1-h3), held by Soul
s3 (verdict cultmath-tapes:verdict:cut-cycles-converters.s3) under ruling
cycles-kernel-reference-ruled. Main carried only docs since the branch base
b3bd8dd5, so the merged code is the Soul-verified head 53f4c213.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 05:57:52 +02:00
Meta Vens
f1ee728d26 cultmath-tapes map: Pass 7 (tape unparked for trophic graphs: body facts T1-T7, model page, cut sequence); tape target status
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 05:57:41 +02:00
Meta Vens
9ec66d77f7 cultlib-gaps map: section D r2 (D10-D14: stack between checks scales with record size; door-top check, 512 KiB red zone, 16 KiB record bound)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 05:00:47 +02:00
Meta Vens
53f4c21311 CultMath: infinite.json written by oracle.py on Yggdrasil (Blender 5.2.2, SSE4.2); CultMath.Cycles.glsl regenerated for the power parity
infinite.json holds 47 cases of infinite operands for cycles_power, cycles_modulo and Map Range Smooth Step, made inside Blender's kernel; only that family was regenerated. CultMath.Cycles.glsl changed by the generator alone (CULTMATH_WRITE_GLSL=1); glsl-parity.json and CultMath.glsl are byte-identical. This commit builds and its tests are expected green.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 04:46:11 +02:00
Meta Vens
77a60d2910 CultMath Cycles: an infinite exponent is even in cycles_power; the oracle's infinite family and the pins for infinite operands
cycles_power read b - 2 floor(0.5 b) != 0 as odd, which is NaN for b = +-inf, so a negative base got a negative result where Cycles' powf(a, +-inf) is never negative. Odd now means the remainder is exactly 1 (C#, HLSL, Unity copy). oracle.py gains the infinite family (operands made in-kernel as +-exp(100), results read through comparisons with 0) for power, modulo and Map Range Smooth Step; the infinite.json fixture and the regenerated GLSL follow in the next commit. This commit builds; InfiniteOperandsFollowBlender and CommittedGlslEqualsLowering fail until it lands.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 04:42:39 +02:00
Meta Vens
6c902fd449 cultlib-gaps map: P17 records the mixed-width NaN key refusal as a deliberate exception (py-msgpack-default r5)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 03:19:03 +02:00
Meta Vens
d8f38057a1 CultMath: the six cycles-oracle fixtures regenerated by oracle.py on Yggdrasil (per-function seeds, arcsine/arccosine/arctan2 rows, overflowing smoothstep rows); CultMath.Cycles.glsl and glsl-parity.json regenerated
Fixtures are the oracle's output (SSE4.2 kernel, Blender 5.2.2 Linux), two runs cmp-identical; CultMath.glsl line 1 (the
Cycles function list in the generated header) carries the three new names.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 03:00:32 +02:00
Meta Vens
c65725da8f CultMath Cycles: arcsine, arccosine and arctan2 with Cycles' guards; per-function oracle seeds; the ruled unfused reference, empty Tolerances and the class table
Fixtures and the regenerated GLSL follow in the next commit, so the tests that read the new rows
(InverseTrigonometry..., OperandStreams...) are red at this commit.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 02:45:51 +02:00
Meta Vens
4dd19467c9 cultmath-tapes map: the Cycles cuts under cycles-kernel-reference-ruled (AVX2 single-rounding probe, documented bound)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 02:38:35 +02:00
Meta Vens
b6118cfd45 cultlib-gaps map: section D (decode-stack-safety probes D1-D9, authority) and cut-order item 7
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 02:36:15 +02:00
Meta Vens
beb8362fbc cultlib-gaps map: py-msgpack-default r4 (convert to a new file under py-migration-shape-ruled; P14-P17)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-14 02:01:36 +02:00