VonNeumannCore: The Stored-Program Form, Verified in Lean 4 — An Interpreter Proved to Simulate Its Machine

In the stored-program architecture, instructions are data: a program resides in memory, and a single fetch–decode–execute interpreter runs whatever the memory holds. This paper presents VonNeumannCore, a Lean 4 development that turns this architecture into proved statements about a universal register machine. Each instruction is encoded into three memory cells (an opcode and two arguments); the encoding is proved faithful, in that decoding is its inverse (decode_encode) and the canonical memory image of a program decodes back to that program at every instruction (encodeMem_encodes). The interpreter, which reads its instruction from memory at each cycle and never writes to memory, is proved to simulate the directly executed machine step for step (vsim, vsim_iter): at every step count, interpreted and direct execution agree on the program counter and the registers, with the memory untouched. The machine's halting guarantee is then proved to transfer to interpreted execution (interpreter_halts), and the worked example is run from memory to the same proved final state as its direct execution (progZero_interpreted). The development is self-contained in core Lean, without Mathlib and without sorry; its checked axiom basis is [propext, Quot.sound].

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24215472
Citations
1
Primary Topic
Logic, programming, and type systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

VonNeumannCore: The Stored-Program Form, Verified in Lean 4 — An Interpreter Proved to Simulate Its Machine

Takeo Yamamoto
1 citations
Zenodo (CERN European Organization for Nuclear Research)
Logic, programming, and type systems
preprint

VonNeumannCore: The Stored-Program Form, Verified in Lean 4 — An Interpreter Proved to Simulate Its Machine

Takeo Yamamoto
preprint en
1 citations

Abstract

In the stored-program architecture, instructions are data: a program resides in memory, and a single fetch–decode–execute interpreter runs whatever the memory holds. This paper presents VonNeumannCore, a Lean 4 development that turns this architecture into proved statements about a universal register machine. Each instruction is encoded into three memory cells (an opcode and two arguments); the encoding is proved faithful, in that decoding is its inverse (decode_encode) and the canonical memory image of a program decodes back to that program at every instruction (encodeMem_encodes). The interpreter, which reads its instruction from memory at each cycle and never writes to memory, is proved to simulate the directly executed machine step for step (vsim, vsim_iter): at every step count, interpreted and direct execution agree on the program counter and the registers, with the memory untouched. The machine's halting guarantee is then proved to transfer to interpreted execution (interpreter_halts), and the worked example is run from memory to the same proved final state as its direct execution (progZero_interpreted). The development is self-contained in core Lean, without Mathlib and without sorry; its checked axiom basis is [propext, Quot.sound].

Zenodo (CERN European Organization for Nuclear Research)
Logic, programming, and type systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon