Deterministic Invariant Systems Research Laboratory
Central documentation and coordination hub for the RFT-SIRM ecosystem
| Audience | Document | What You Will Learn |
|---|---|---|
| 🎯 First-time visitor | This README | What UltraCore is, why it exists, and where everything lives |
| 🌐 Interactive overview | rft-sirm.github.io | Live laboratory website with metrics, case studies, and evidence |
| 🖥️ Live demo (Devnet) | rift-network.vercel.app | Interactive web client for the Solana on-chain protocol, running on Devnet |
| 🏛️ Strategic context | docs/platform.md | Why UltraCore is an execution architecture, not just a blockchain |
| 🔺 Physics research | Phi-Genesis · Live site | Honest, test-driven audit of a fermion mass-formula conjecture via spectral computation on the Sierpiński gasket — claims verified, rejected, or marked open, never just asserted |
| 🔷 Math research | Evgeny-Theorem · Live site | A closed-form gauge-invariant fourth spectral moment for a noncommutative SU(2) connection on the Sierpiński gasket — verified numerically for levels 1–7 to machine precision (55/55 tests), reproducible by anyone |
| 🤖 AI / LLM | AI_GUIDE.md | How to interpret research vs. engineering, metaphors vs. claims |
| 🏗️ Engineer | ARCHITECT.md | System design, components, and technical decisions |
| 🔬 Researcher | SCIENTIFIC_BASIS.md | Disciplinary foundations and methodological boundaries |
| 💼 Investor / Partner | PITCH.md | Full dossier with metrics, roadmap, and evidence |
One-sentence summary: UltraCore RFT is a research laboratory building a deterministic execution platform for distributed systems, where mathematical invariants are hard constraints enforced after every state transition.
flowchart TB
subgraph MATH["Mathematical Core"]
I1["I1: Supply Conservation"]
I2["I2: Mint/Burn Accounting"]
I3["I3: Dust Bound"]
I4["I4: Debt Limit"]
end
subgraph RUNTIME["Runtime Layer"]
MEM["Memory Contexts<br/>CPI Isolation"]
SCHED["Conflict-Aware<br/>Scheduler"]
end
subgraph VERIFY["Verification"]
FUZZ["4.29B+ Fuzz<br/>Executions"]
SEL4["seL4 CDT<br/>1B+ Ops"]
AUDIT["14 Findings<br/>Addressed"]
end
MATH --> RUNTIME
MATH --> VERIFY
RUNTIME --> VERIFY
| Metric | Value |
|---|---|
| Fuzz Executions | 4.29B+ |
| Invariant Violations | 0 |
| Security Findings Fixed | 14 |
| Upstream RFCs | 2 |
| seL4 Kernel Crashes | 0 |
| Daily CI Fuzzing | 5h 55m |
UltraCore RFT is best understood as an execution architecture — a deterministic execution substrate — rather than as a single blockchain or mathematical theory.
flowchart TB
subgraph SF["Scientific Foundations"]
MATH["Mathematics · Graph Theory · Category Theory<br/>Information Theory · Dynamical Systems · Invariant Theory"]
end
subgraph CONCEPT["Conceptual Framework"]
RFT["Reality Fractal Theory"]
end
subgraph MODEL["Execution Model"]
SIRM["Stable Invariant Rift Model<br/>I1–I4 Hard Constraints"]
end
subgraph RUNTIME["Runtime Engine"]
UC["UltraCore Runtime<br/>Deterministic State Machine · O(1) Distribution"]
end
subgraph ARCH["Execution Architecture"]
MEM["Memory Contexts · CPI Isolation"]
SCHED["Conflict-Aware Scheduler"]
ROLL["Permission Rollback · Atomic Operations"]
end
subgraph BLOCK["Blockchain Layer"]
L1["Rift-L1 Blockchain<br/>Standalone Validator Core"]
NET["Rift Network<br/>Solana On-Chain Protocol"]
end
subgraph FUTURE["Future Targets"]
HA["High-Assurance Infrastructure"]
CRIT["Critical Computing Environments"]
KERNEL["Additional Verified Kernels"]
end
SF --> CONCEPT
CONCEPT --> MODEL
MODEL --> RUNTIME
RUNTIME --> ARCH
ARCH --> BLOCK
BLOCK --> FUTURE
ARCH -.->|"kernel-agnostic"| FUTURE
Key insight: The blockchain is one implementation. The runtime is another. The verification methodology is another. Together they form one coherent architecture — layered, verifiable, and kernel-agnostic.
See docs/platform.md for the full strategic identity document.
SIRM = Stable Invariant Rift Model. It is the mathematical core of every RFT-SIRM system.
All systems enforce four hard constraints after every state-mutating operation:
I1: total_supply = total_base_sum + global_field * p
I2: total_supply = total_minted - total_burned
I3: dust_accumulator < p (when p > 0)
I4: effective_balance[i] >= -(total_supply / 10p)
Where effective_balance[i] = base_balance[i] + global_field.
This model enables O(1) distribution: updating global_field by a scalar delta changes every participant's effective balance simultaneously, regardless of participant count. No iteration. No per-account writes.
See docs/foundations.md for the mathematical derivation.
| Repository | Role | Status | Key Evidence |
|---|---|---|---|
| Rift-L1-Blockchain | Standalone L1 runtime | Active | 1T+ ops, 0 invariant violations |
| Rift-Network | Solana on-chain protocol | Audited | 14 findings addressed, 2.5B+ fuzz runs · Live Demo (Devnet) |
| agave-abiv2-memory-contexts | SVM memory isolation (PoC) | Research Complete | 4.29B+ exec, PoC-only bug found & documented — upstream uses abi_v2_prepare_for_instruction architecture |
| agave-rift-scheduler | Conflict-aware scheduling | Active | 91M exec/run, agave#14274 |
| aave-v4-hub-model-review | DeFi ledger invariant model (Aave V4 Hub) | Complete | 184K ops, 0 violations, complementary to Certora FV |
| research/seL4 | Kernel verification | Complete | 1B+ ops deterministic fuzzing |
| Phi-Genesis | Fractal spectral physics — mass-formula audit | Active | 14/14 tests passing · 2 claims formally rejected (η-invariant, ad hoc topology fit) · 2 open problems documented, not hidden · live site |
| Evgeny-Theorem | Noncommutative spectral geometry — SU(2) gauge theory on fractals | Active | 55/55 tests passing · closed-form H⁴ identity verified to 1e-13, held-out cross-check, gauge invariance to 8e-15 · live site |
An exact, closed-form, gauge-invariant fingerprint of non-commutativity for an SU(2) gauge field on a fractal — one number at every refinement level, reproducible in seconds.
The object. The Sierpiński-gasket graph SG(m) with Hilbert space C^n ⊗ C² and an SU(2)-valued connection whose rotation axis cycles x / y / z from triangle to triangle, so that the holonomies of neighbouring triangles genuinely do not commute. It is compared against the commuting control C′ (fixed axis: exactly two decoupled U(1) copies) at the same angle θ.
The result. The raw fourth-moment trace defect is given exactly by
The invariant. Normalizing by dim(H) = 3^(m+1) + 3 gives the intensive quantity
Equivalently, with F = 3^m triangular faces, the defect is linear in the face count:
| m | dim(H) | Δ_m(H⁴, π/2) | I_m(π/2) | distance to −8/9 |
|---|---|---|---|---|
| 1 | 12 | -16 | -1.333333 | 0.444444 |
| 2 | 30 | -32 | -1.066667 | 0.177778 |
| 3 | 84 | -80 | -0.952381 | 0.063492 |
| 4 | 246 | -224 | -0.910569 | 0.021680 |
| 5 | 732 | -656 | -0.896175 | 0.007286 |
| 6 | 2,190 | -1,952 | -0.891324 | 0.002435 |
| 7 | 6,564 | -5,840 | -0.889701 | 0.000813 |
| Property | Evidence | |
|---|---|---|
| 🎯 | Exact | One closed form gives Δ_m for any m and θ in O(1). At level 20 the operator has more than 10¹⁰ dimensions; the closed form costs a few arithmetic operations. (Applies to this single quantity only, not to the full spectrum.) |
| 🔒 | Gauge-invariant | Under a Haar-random SU(2) gauge transformation the spectrum changes by at most 8 × 10⁻¹⁵ and M₄ is identical |
| 🧪 | Held-out tested | Five (m, θ) pairs chosen after the formula was fixed agree to 10⁻¹¹–10⁻¹⁴; a 20-point θ-grid agrees to below 10⁻⁹ |
| 📐 | A non-Abelian witness | The defect is exactly zero for moments p = 1, 2, 3 and first appears at p = 4; it vanishes in the commuting limit |
| ♻️ | Reproducible | 55 fast checks run in seconds; level 7 (dim 6,564) agrees with the closed form to 9 × 10⁻¹³ absolute, 1.6 × 10⁻¹⁶ relative |
| Statement | Status |
|---|---|
Closed form equals direct computation for every tested (m, θ), m = 1…7 |
✅ Established (reproducible) |
| Gauge invariance of the fourth moment | ✅ Established |
The closed form holds for all m and θ |
🔬 Verified numerically — analytic proof in progress |
| Lean 4 formalization | 🔬 Scaffold only; the operator-level identity is not yet machine-checked |
| Peer review / independent replication | 📅 Not yet |
- Quantum simulation: a known-answer benchmark for simulators of non-Abelian gauge dynamics. The state space is a site register plus one spin-½: level 7 fits in 13 qubits by dimension count. Circuit cost is not yet studied.
- Trace estimation: an exact test vector for spectral-moment estimators, quantum or randomized classical.
- Further identities: the same framework may yield exact results for other moments, fractals and gauge groups.
Not claimed: any speed-up for quantum algorithms, any statement about physical gauge theories, or any result on Millennium Prize Problems. Details, validation program and boundaries: PITCH.md.
git clone https://github.com/RFT-SIRM/Evgeny-Theorem.git && cd Evgeny-Theorem
python3 -m venv .venv && source .venv/bin/activate
pip install -r reproducibility/requirements.txt
pytest tests/ -m "not slow" -v # 55 checks, seconds
pytest tests/ -m slow -v # level 7, a few minutesFull statement: THEOREM.md · Numerical record: VERIFICATION.md · Mathematical framework: docs/foundations.md
Every claim is backed by reproducible verification. We measure correctness rather than asserting it.
| Layer | Method | Evidence |
|---|---|---|
| L1 — Static | Clippy, Miri, cargo-audit | Every push |
| L2 — Engineering | Unit + integration + differential tests | 15+ tests per component |
| L3 — Fuzzing | libFuzzer deterministic fuzzing | 4.29B+ exec, 0 invariant violations |
| L3b — Kernel | seL4 CDT complementary verification | 1B+ ops, 0 kernel crashes |
| L3c — DeFi Model | Python deterministic state-machine fuzz | 184K ops, 0 INV violations |
| L4 — Formal | TLA+ / Coq | Planned |
See docs/field_trials.md for the full verification report.
Independent engineering validation of the formally verified seL4 microkernel:
- Subsystem: Capability Derivation Tree (CDT)
- Operations: > 1.0 × 10⁹
- Kernel crashes: 0
- Post-marathon test suite: 123 / 123 passed
Important: This was infrastructure research, not a claim of production deployment. See SEL4_CDT_FUZZING.md and docs/field_trials_sel4.md.
Full documentation is built with MkDocs Material:
pip install -r requirements.txt
mkdocs serve| Document | Description | Audience |
|---|---|---|
| docs/platform.md | Strategic identity: what UltraCore is and why it matters | Everyone |
| docs/architecture.md | Detailed architecture with Mermaid diagrams | Engineers |
| docs/foundations.md | Formalized SIRM invariants | Researchers |
| docs/field_trials.md | Verification results & readiness checklist | Validators |
| docs/field_trials_sel4.md | seL4 CDT stress-verification report | OS Researchers |
| docs/strategy.md | Full development strategy | All |
| docs/implementation.md | Build instructions and component architecture | Developers |
| docs/glossary.md | Terminology and definitions | All |
| docs/support.md | Research support and collaboration | All |
See CONTRIBUTING.md and CODE_OF_CONDUCT.md. For security disclosures, see SECURITY.md.
Copyright 2026 Eugeny (RFT-SIRM). Licensed under Apache 2.0.