Monitor CPU and GPU usage while running heavy workloads like machine learning or video editing on a Mac
Check real-time power consumption and clock speeds to understand performance during demanding tasks
Track memory and swap usage to see if your Mac is running out of RAM under load
Must be run with sudo because it uses macOS powermetrics, which requires administrator access.
asitop is a command-line performance monitoring tool for Apple Silicon Macs (the M1 and later chips). It shows live stats about how hard the processor, graphics, and memory are working, displayed in the terminal. The author built it because nothing similar existed for Apple Silicon at the time. The tool displays CPU utilization split by efficiency cores and performance cores, GPU utilization, current clock frequencies for both, memory usage including RAM and swap space, and power consumption for the CPU and GPU. It also shows a running chart of power draw and a rolling average, so you can see spikes over time rather than just a single instant reading. Some figures like maximum power limits are hardcoded estimates because Apple does not expose them through official channels on the system. Under the hood, asitop reads data from several macOS utilities. The main one is powermetrics, a built-in macOS tool that taps hardware performance counters. Because powermetrics needs administrator access, asitop must be run with sudo. It also pulls memory and swap stats from a Python library called psutil, reads CPU and core counts from sysctl, and checks GPU core count via system_profiler. Installation is done with pip, the standard Python package manager: pip install asitop. Once installed, you run it in the terminal with sudo asitop. Optional flags let you adjust how often the display refreshes, change the color theme, and set the window used for averaged readings. The README notes it only works on Apple Silicon Macs running macOS Monterey or later. asitop is lightweight by design and is meant to have minimal impact on the machine it is monitoring. It is released as an open Python project with no license specified in the README.
← tlkh on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.