The Secret Apple M1 AMX: Apple Matrix Coprocessor #ReverseEngineering #M1 #Apple @Apple @erikengheim

Erik Engheim posts on Medium about Developer Dougall Johnson thoroughly reverse engineering the Apple M1 processor, uncovering a secret powerful coprocessor dubbed AMX: Apple Matrix coprocessor.

A matrix is basically just a table of numbers. If you have worked with spreadsheets such as Microsoft Excel, you have basically worked with something very similar to matrices. The key difference is that in math such tables of numbers have a laundry list of operations they support and specific behavior.

The reason matrices are important is because they are heavily used in:

  • Image processing
  • Machine learning
  • Speech and handwriting recognition
  • Face recognition
  • Compression
  • Multimedia: audio and video

In particular machine learning which has been hot these last years. Just adding more cores to the CPU will not make this run fast enough as it is really demanding. You really need specialized hardware. Regular tasks such browsing the internet, writing email, word processing and spreadsheets has been running fast enough for years. It is for specialized tasks which we really need to boost the processing power.

Why Is the AMX a Secret?

Yet Apple has indeed extended their ARM CPU cores with AMX instructions. How do we know? Apple has kept this a secret. There is no manual, publicly available, describing these instructions. However developer Dougall Johnson has done an amazing reverse engineering effort on the M1 to discover this coprocessor. His efforts are described here. For matrix related math operations Apple has special libraries or frameworks such as Accelerate, which is made up of:

  • vImage — higher level image processing, such as converting between formats, image manipulation.
  • BLAS — a sort of industry standard for linear algebra (what we call the math dealing with matrices and vectors).
  • BNNS — is used for running neural networks and training.
  • vDSP  digital signal processing. Fourier transformations, convolution. These are mathematical operations important in image processing or any signal really including audio.
  • LAPACK — higher level linear algebra functions, e.g. for solving linear equations.

Johnson wrote special programs to analyze and observe what these programs did to discover the special undocumented AMX machine code instructions.

Read Engheim’s analysis here.


Stop breadboarding and soldering – start making immediately! Adafruit’s Circuit Playground is jam-packed with LEDs, sensors, buttons, alligator clip pads and more. Build projects with Circuit Playground in a few minutes with the drag-and-drop MakeCode programming site, learn computer science using the CS Discoveries class on code.org, jump into CircuitPython to learn Python and hardware together, TinyGO, or even use the Arduino IDE. Circuit Playground Express is the newest and best Circuit Playground board, with support for CircuitPython, MakeCode, and Arduino. It has a powerful processor, 10 NeoPixels, mini speaker, InfraRed receive and transmit, two buttons, a switch, 14 alligator clip pads, and lots of sensors: capacitive touch, IR proximity, temperature, light, motion and sound. A whole wide world of electronics and coding is waiting for you, and it fits in the palm of your hand.

Join 30,000+ makers on Adafruit’s Discord channels and be part of the community! http://adafru.it/discord

Have an amazing project to share? The Electronics Show and Tell is every Wednesday at 7pm ET! To join, head over to YouTube and check out the show’s live chat – we’ll post the link there.

Join us every Wednesday night at 8pm ET for Ask an Engineer!

Follow Adafruit on Instagram for top secret new products, behinds the scenes and more https://www.instagram.com/adafruit/

CircuitPython – The easiest way to program microcontrollers – CircuitPython.org


Maker Business — Pololu’s account of the chip shortage

Wearables — Getting overly ambitious with EL wire

Electronics — Invisible Bugs

Python for Microcontrollers — Python on Microcontrollers Newsletter: MicroPython SAMD Support, ESP32-S3 Feather and much more! #Python #Adafruit #CircuitPython @micropython @ThePSF

Adafruit IoT Monthly — Energy Harvesting Sensor Nodes, Tracking a Turtle, , and more!

Microsoft MakeCode — MakeCode Thank You!

EYE on NPI — Maxim’s Himalaya uSLIC Step-Down Power Module #EyeOnNPI @maximintegrated @digikey

New Products – Adafruit Industries – Makers, hackers, artists, designers and engineers! — New Prods 11/17/21 feat. Adafruit KB2040 – RP2040 Kee Boar Driver!

Get the only spam-free daily newsletter about wearables, running a "maker business", electronic tips and more! Subscribe at AdafruitDaily.com !



No Comments

No comments yet.

Sorry, the comment form is closed at this time.