How to % CPU

A common question among sysadmins: "How do I quickly generate dummy CPU load on this machine?" If a simple 100% CPU load across one or more cores is sufficient, building a custom solution is surprisingly easy.

The Single-Core Solution (One-Liner)

This single line of C code, compiled and run, will peg one core at 100% utilization:

int main() {while (1) {}}

Compile using gcc -o stressme stressme.c (or cl stressme.c on Windows) and run with ./stressme (or stressme.exe). To stress multiple cores, simply run multiple instances of the program.

Multi-Core Solution (Multi-Threaded)

For more controlled multi-core stress testing, consider this multi-threaded version using 4 POSIX threads:

#include <pthread.h>
#include <unistd.h>

#define NUM_THREADS 4

void *loop(void *arg) {
    while (1) {}
}

int main() {
    pthread_t threads[NUM_THREADS];
    for (int i = 0; i < NUM_THREADS; i++) {
        pthread_create(&threads[i], NULL, loop, NULL);
    }
    for (int i = 0; i < NUM_THREADS; i++) {
        pthread_join(threads[i], NULL);
    }
    return 0;
}

Compile with gcc -o multistress multistress.c -pthread.

How It Works

The core functionality is a simple infinite loop (while (1) {}). At the assembly level, this translates to a continuous jmp instruction, maximizing CPU usage. Modern preemptive multitasking operating systems allow this loop to consume processor time without rendering the system completely unresponsive; the process can still be terminated. In contrast, older cooperative multitasking systems would likely freeze due to such a loop.

global _start

_start:
    jmp _start

The above is the detailed content of How to % CPU. For more information, please follow other related articles on the PHP Chinese website!