overflow-test.cpp File Reference

Standalone microbenchmark and overflow-trip harness for the safe:: primitives in overflow.hpp. More...

#include <stdio.h>
#include "overflow.hpp"
#include "assert.h"
#include <stdlib.h>
#include <string.h>

Go to the source code of this file.

Macros
#define	outer   20000
#define	inner   5000
#define	stmt   x = 1
#define	roll   100

Functions
int	main (int argc, char **argv)

Variables
volatile int	s0 = 100
volatile int	x = 200

Detailed Description

Standalone microbenchmark and overflow-trip harness for the safe:: primitives in overflow.hpp.

Note

AI-generated documentation. Verify against the source before relying on it.

Not part of the engine library — a separately-compiled binary (make DEPENDS=no overflow-test) developers reach for to confirm overflow detection and measure the per-call cost of safe::add, safe::mult, safe::sub, safe::div, safe::minus, and safe::ov (sub-nanosecond-per-iteration on modern hardware). main has two modes: with a command-line argument in {sub, add, mult, div, minus, throw} it deliberately passes overflowing operands (safe::add(INT_MAX, 1), etc.) and exits with the expected fatal error, so the developer can verify each detector individually; with no argument it runs a hot loop of outer * inner * roll = 20000 * 5000 * 100 = 10^10 iterations of a chosen stmt macro — the preserved #if 0 block holds the developer's previous timing comparisons (y = j + x vs y = safe::add(j, x) at 10^10 reps, etc.) and the bare-x = 1 loop-overhead measurements (single assignment 4.488s for 5*10^9 reps ≈ 0.9ns each, pair 8.973s ≈ 1.8ns per pair) that quantify how much of the per-iteration cost is overflow checking versus loop overhead.

Build with optimisation but no LTO so call sites stay distinct; no engine state is involved.

See also

overflow.hpp

Definition in file overflow-test.cpp.