Historical report — qualified October 27, 2026. The original body is preserved as a dated description, not a current implementation or security certification. Claims of production readiness, universal determinism, complete integration, or capability inferred solely from RFCs require source and test evidence. Use the current research overview, surface register, and alignment audit for present interpretation.
Generated: April 4, 2026
Scope: Complete architectural overview of T81 Foundation systems and subsystems
Status: Production-grade analysis with experimental frontier documentation
T81 Foundation is a deterministic, policy-gated runtime stack for auditable computation. The system is organized into two distinct domains:
The architecture maintains strict boundaries between stable production guarantees and experimental research, ensuring system integrity while enabling innovation.
isa/, include/t81/isa/vm/, include/t81/vm/vm.cpp - Main interpreter dispatch (301KB)tensor_helpers.cpp - Tensor operations (64KB)memory_segments.cpp - Memory management (31KB)policy_trace_bridge.cpp - Axion integration (9KB)kernel/axion/, include/t81/axion/fs/, include/t81/canonfs/lang/, include/t81/lang/lang/c_frontend/)lang/mlir/)lang/llvm/)vm/
├── vm.cpp # Main interpreter loop (301KB)
├── decoder.cpp # Instruction decoding (63KB)
├── decode_state.cpp # State management (56KB)
├── tensor_helpers.cpp # Tensor operations (64KB)
├── memory_segments.cpp # Memory management (31KB)
├── dpe/ # Deterministic Parallel Execution
├── determinism/ # Determinism verification
└── jit/ # Just-in-time compilation
kernel/axion/
├── policies/ # Policy definitions
├── shell/ # Shell integration
└── userenv/ # User environment
fs/
├── canonfs_interchange.cpp # Import/export (36KB)
├── auto_optimizer.cpp # Performance optimization (16KB)
├── ml_optimizer.cpp # ML-based optimization (23KB)
├── persistent_driver.cpp # Persistent storage (11KB)
└── performance_analyzer.cpp # Performance analysis (12KB)
tools/cli/
├── core/ # Essential CLI commands
├── ai_tools/ # AI and inference tools
├── canonfs_tools/ # CanonFS utilities
├── bundle_tools/ # Bundle management
└── experimental/ # Experimental tools
tests/
├── cpp/ # 393 core tests (all passing)
├── determinism/ # Determinism verification
├── ci/ # Continuous integration
└── corpus/ # Test data
cmake/
├── Compiler.cmake # Compiler configuration
├── Dependencies.cmake # Dependency management
├── Options.cmake # Build options
└── Platform.cmake # Platform detection
experimental/
├── ai/ # Advanced AI concepts (47 items)
│ ├── csi/ # Controlled Stochastic Inference
│ ├── cognitive/ # Cognitive architectures
│ └── governance/ # AI governance frameworks
├── cog/ # Cognitive tier architectures
├── tiers/ # Multi-level systems (14 items)
├── ai_governance/ # Multi-policy frameworks
└── ai_integration_demos/ # Proof-of-concept demos
experimental/
├── distributed/ # Distributed systems
├── global_infrastructure_concepts/ # Global bundle concepts
├── marketplace_prototypes/ # Economic systems
└── hanoi/ # Advanced VM architectures
experimental/
├── setun/ # Ternary computing experiments
├── benchmark_prototypes/ # Experimental measurements
├── marketing_arguments/ # Positioning materials
└── ai_governance/ # Governance frameworks
include/t81/
├── t81.hpp # Main API header
├── canonfs.hpp # CanonFS interface
├── vm/ # VM APIs
├── axion/ # Policy APIs
├── tensor.hpp # Tensor operations (21KB)
└── types/ # Type definitions (60 items)
ffi/
├── c_api/ # C API bindings
├── python/ # Python bindings
└── ffi_dispatcher.cpp # FFI dispatch (15KB)
include/t81/
├── c_frontend/ # C language support
├── python_frontend/ # Python language support
├── rust_frontend/ # Rust language support
└── llvm/ # LLVM integration
codec/
├── base81.cpp # Base-81 encoding
├── base243.cpp # Base-243 encoding
└── advanced_ternary_quantization.cpp # Quantization
inference/
├── batch_inference_engine.cpp # Batch processing
├── simd/ # SIMD optimizations
└── tensor/ # Tensor operations
isa/math/
├── quantization/ # Quantization algorithms
└── t81_soft_math/ # Software math library
The DCP defines the guaranteed deterministic surface:
Experimental subsystems are strictly separated:
benchmarks/
├── BM_Base81SIMD.cpp # Base-81 SIMD benchmarks
├── BM_CanonFS.cpp # CanonFS performance
├── BM_DPE.cpp # Deterministic Parallel Execution
└── results/ # Benchmark results archive
fs/
├── auto_optimizer.cpp # Automatic optimization
├── ml_optimizer.cpp # ML-based optimization
├── performance_analyzer.cpp # Performance analysis
└── ternary_optimization.cpp # Ternary-specific optimizations
inference/simd/
├── base81_digits_avx2.cpp # AVX2 optimizations
└── include/t81/simd/ # SIMD intrinsics
docs/architecture/
├── OVERVIEW.md # System architecture overview
├── adr/ # Architecture Decision Records
├── diagrams/ # Architecture diagrams
└── cross-cutting/ # Cross-cutting concerns
spec/
├── rfcs/ # 120+ RFC documents
├── conformance/ # Conformance specifications
├── supplemental/ # Supplementary specifications
└── companion/ # Companion specifications
examples/
├── ai-and-inference/ # AI integration examples
├── canonfs-enhanced-errors-demo/ # Error handling demos
├── compiler-and-ffi/ # Compiler and FFI examples
└── aspirational/ # Future concept demos
crypto/
├── sha3.cpp # SHA-3 implementation
└── include/t81/crypto/ # Crypto APIs
kernel/axion/policies/
├── Policy definitions # Governance rules
├── Enforcement mechanisms # Policy implementation
└── Audit trails # Compliance tracking
vm/determinism/
├── Determinism validation # Determinism checking
└── Evidence collection # Provenance tracking
market-demos/
├── financial_services/ # HFT trading demo
├── healthcare/ # Medical diagnosis demo
├── legal_services/ # Document analysis demo
├── industrial/ # Safety monitoring demo
└── customer_demo_system.cpp # Unified demo system
customer_tools/
├── roi_calculators/ # Financial analysis tools
└── engagement_tools/ # Customer engagement utilities
strategic_docs/
├── CUSTOMER_ACQUISITION_STRATEGY.md
├── CUSTOMER_ENGAGEMENT_STRATEGY.md
└── STRATEGIC_EXECUTION_PLAN.md
.github/workflows/
├── ai-experiments-ci.yml # AI experiments CI
├── bench.yml # Benchmark CI
└── 15+ other workflows # Comprehensive CI/CD
execution_reports/
├── VERIFICATION_REPORT.md # Performance verification
├── WEEK2_EXECUTION_SUMMARY.md
├── WEEK3_EXECUTION_SUMMARY.md
└── WEEK4_EXECUTION_SUMMARY.md
scripts/governance/
├── c2_month_close_check.py # Monthly close validation
├── c2_month_close_preflight.py
└── 22+ governance scripts # Comprehensive governance
The system maintains strict architectural boundaries:
T81 implements a three-layer intelligence model:
Layer 0: Deterministic Substrate (DCP)
├── TISC ISA, T81VM, Axion, CanonFS
└── Bit-exact reproducibility guaranteed
Layer 1: Governed Stochastic Processes (CSI)
├── Controlled Stochastic Inference
├── Policy-gated sampling with provenance
└── Accountable uncertainty
Layer 2: Unbounded External AI
├── External model integration
└── Research-only exploration
The missing subsystem for accountable uncertainty:
T81 supports incremental adoption through:
The architecture enables evolution through:
T81 Foundation represents a comprehensive deterministic computing platform with clear architectural boundaries between production guarantees and experimental research. The system’s strength lies in its deterministic core profile while maintaining an innovative experimental frontier for future development.
The architecture successfully balances:
This systems and subsystems report provides a complete overview of the T81 Foundation architecture, enabling informed decisions about adoption, contribution, and research directions.
Report Status: Complete and Current
Next Review: Align with major releases or architectural changes
Contact: Refer to project documentation for contribution guidelines