DIG: Dynamic Invariant Generator

DIG infers program invariants — nonlinear/linear equalities, octagonal inequalities, congruences, and min/max-plus relations — from program code or execution traces. [GitHub] [ROARS Lab]


1. Program input

Load example:

Input type:  

2. Options

(optional — click to configure DIG; defaults are used otherwise)
Numeric parameters (leave blank for DIG defaults)
find nonlinear invariants up to this degree
autodegree (max number of terms)
degree for inequalities (1 = linear, 2 = quadratic, ...)
number of terms for inequalities (2 = octagonal)
coefficients for inequalities (1 = [-1,0,1], i.e., oct)
max upperbound value for inequalities
depth limit of symbolic execution
number of random inputs (only with -noss)
max input value
random seed (for reproducibility)
logger verbosity (0 = error ... 4 = detail)
wall-clock limit enforced by the web runner (max 300)
Disable / debugging flags









Terms / LLM user-supplied terms, separated by ;

max LLM propose/verify rounds

  Ready.


3. Discovered invariants

(no results yet)

4. Raw execution output

(waiting for execution output)

DIG · ROARS Lab · George Mason University