🍎 apple-smi
nvidia-smi equivalent for Apple Silicon – High-performance, sudoless GPU monitoring with zero dependencies.
$ apple-smi
Wed Mar 04 14:52:00 2026
+-----------------------------------------------------------------------------------------+
| APPLE-SMI 0.1.4 macOS Version: 26.3 (25D125) Metal Version: 3 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name | Disp.A | |
| Temp Pwr:Usage/Cap | Memory-Usage | GPU-Util |
|=========================================+========================+======================|
| 0 Apple M1 (8-Core GPU) | On | |
| 30C 2.4W / 20W | 13826MiB / 16384MiB | 12% |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU PID Type Process name GPU Memory |
| Usage |
|=========================================================================================|
| 0 72743 C python3.11 530MiB |
| 0 1007 G NotificationCenter 170MiB |
| 0 3197 G WeType 216MiB |
+-----------------------------------------------------------------------------------------+
✨ Features
- 🚫 No sudo required – Uses private macOS APIs (
IOReport,IOHIDSensors,SMC) for sudoless operation. - ⚡ Ultra-Fast & Lightweight – Optimized startup (<50ms) using targeted
sysctlandioregcalls. - 🔋 Power Monitoring – Displays SoC power consumption (Usage) against chip TDP (Cap).
- 🌡️ Smart Thermals – High-accuracy GPU/SOC temperature readings with sensor fallback logic.
- 📋 Robust Process Listing – Lists active GPU processes with full name resolution (no truncation for paths with spaces).
- 🔍 C/G Classification – Categorizes processes by Type (
Cfor Compute,Gfor Graphics). - 💾 Memory Usage – Detailed breakdown of unified memory utilization.
- 🔄 Watch Compatible – Default 100ms sampling interval for smooth real-time monitoring.
- 📦 Zero Dependencies – Pure Python, utilizing
ctypesto interface with macOS frameworks.
📥 Installation
pip install apple-smi
🚀 Usage
Command Line Options
# Basic usage
apple-smi
# Set sampling interval (default: 100ms)
apple-smi --interval 250
# JSON output for integration with other tools
apple-smi --json
# Continuous monitoring (highly recommended)
watch -n 0.1 apple-smi
Process Filtering
By default, apple-smi only shows Compute (C) processes (e.g., Python, MLX, Torch, llama.cpp) to reduce system noise. You can show all GPU-connected processes (including window compositors and browsers) using an environment variable:
APPLE_SMI_SHOW_ALL_PROCESSES=1 apple-smi
Backends
apple-smi automatically selects the best available backend:
- IOKit (Default): Uses undocumented IOKit APIs for sudoless monitoring.
- Powermetrics (Fallback): Uses the system
powermetricstool (requiressudoor root privileges).
| Variable | Values | Description |
|---|---|---|
APPLE_SMI_BACKEND |
iokit, powermetrics |
Force a specific backend. Default: auto-detect. |
APPLE_SMI_SHOW_ALL_PROCESSES |
0, 1 |
Show all GPU processes (G+C). Default: 0. |
🔧 Requirements
- macOS on Apple Silicon (M1, M2, M3, M4, M5)
- Python 3.10+
📝 License
MIT License – see LICENSE for details.
Metadata
Release files for apple-smi 0.1.4
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| apple_smi-0.1.4.tar.gz | 26.5 kB | Without provenance | Details for apple_smi-0.1.4.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| apple_smi-0.1.4-py3-none-any.whl | Python 3 | none | any | Details for apple_smi-0.1.4-py3-none-any.whl |
Total release size: 53.2 kB
Release files / apple_smi-0.1.4.tar.gz
| Download URL | apple_smi-0.1.4.tar.gz |
|---|---|
| Size | 26.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e6a787cd5e705241fc4fa8847ad6198c41e559ceac51a5b6c8bc9dacf55583ee
|
|
BLAKE2b-256 checksum How to use checksums |
ff618a882e8e0519d8eb3fb540204f7d88929465c28c2a5e2aac30868a674fa6
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Release files / apple_smi-0.1.4-py3-none-any.whl
| Download URL | apple_smi-0.1.4-py3-none-any.whl |
|---|---|
| Size | 26.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
4e854523435e38fa622ce59c6395a2a158d43842a83211aec1044e86cb3b3e38
|
|
BLAKE2b-256 checksum How to use checksums |
39d169df3d8895d75204c77cff2f80b70918175b68aa93451de6083176452f46
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Security for apple-smi 0.1.4
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.