Mojo Programming Language | Basics

Mojo isn’t publicly available; however, you can access Mojo Programming Language Playground by enrolling in Modular Products. As you sign up, ensure you’ve chosen Mojo in the Modular Product Interest section.

After signing up, you will receive an email granting you access to the Mojo Playground within an hour. In this regard, the Mojo Playground is a JupyterHub environment where users can access the same Mojo standard library. Nonetheless, users also have a private volume to write and save their Mojo programs.

An image from Mojo Playground displays this.

Upon Mojo becoming open-source and publicly available, you can execute a Mojo program from a terminal. It functions as both an interpreted and compiled language.

You have the option to save the file as either test.mojo or test..

Indeed, you’re able to save your file accompanied by the fire emoji. As demonstrated, we’ve crafted a basic Mojo file test. and executed it within the terminal.

Mojo Programming Language and Python: A Comparative Overview

Mojo takes its place as a Python superset, sharing an almost identical syntax. Moreover, it introduces novel attributes—such as let, var, struct, and fn—to establish variables, structures, and functions, augmenting its performance considerably.

The Process of Declaring Variables

When crafting Mojo code, the use of keywords like let and var empowers you to pinpoint variable declarations. This mirrors the approach taken when defining variables in Rust.

Distinct Characteristics of let and var Declarations

The keyword let underscores a variable’s unalterable nature, while var signifies its susceptibility to modification. These designations wield the power to enhance efficiency by enforcing compile-time constraints.

The Role of struct Types

Within Mojo’s framework, the struct keyword emerges, bearing semblance to Python’s class. While Python’s classes entail dynamism and a measured slowness, struct types in Mojo align more with the paradigms of C/C++ and Rust. Their memory layouts, steadfastly fixed at compile time, position them for optimization and native-machine performance.

fn Definitions in Perspective

The act of employing def to outline a function in Python ushers in a conventional Python function, imbued with the anticipated dynamic and adaptable attributes. Contrarily, the fn keyword in Mojo’s context constructs a function of a different caliber—constrained by additional strictures. This signifies the inherent immutability of arguments by default, mandating explicit typing and the declaration of local variables, among other requisites.

Syntax Comparison Example Mojo and Pyhon

Here’s a basic syntax comparison example between Mojo and Python that you can use for your tutorial:

Mojo Programming language code
pyton code

In this example, you can observe the similarities in variable assignment and function definition between Mojo and Python.

In a nutshell, Mojo, by amalgamating features from Python and appending its own, forges a distinct path in the realm of programming languages, offering a blend of familiarity and innovative capabilities.

Matplotlib in Mojo

Mojo Programming Language not only boosts performance but also empowers developers to import any Python library, integrating it with Mojo functions. By making use of the CPython interpreter, Mojo effortlessly caters to all existing Python modules.

In the subsequent illustration, we imported matplotlib.pyplot, and we depicted the line plot. For more intricate examples, explore Mandelbrot in Mojo.

Mojo textcode
Mojo code output

Programming With Mojo Programming Language

Similar to how TypeScript serves as a superset of JavaScript, Mojo extends beyond Python. Alongside introducing new keywords and functions, it remains accessible to any Python programmer, enabling them to comprehend and construct programs with ease.

Within this section, we will look into Mojo’s programming features, highlighting the aspects that contribute to its performance and security.

Low-level programming

Mojo operates as a high-level programming language, enabling access to low-level primitives via the Multi-Level Intermediate Representation (MLIR). This extensible intermediate representation format not only facilitates the implementation of zero-cost abstractions but also capitalizes on robust compiler optimizations.

Tiling optimization and autotune

With a built-in tiling optimization tool, Mojo effectively enhances cache locality and memory access patterns. Achieved by breaking computations into smaller tiles that fit within fast cache memory, this optimization strategy greatly improves overall performance.

The Autotune module within Mojo serves as an interface for adaptive compilation, aiding in the identification of optimal parameters tailored to target hardware. By automating code tuning, it streamlines the process of achieving peak efficiency.

Ownership and borrowing

Mojo employs an ownership and borrowing system to manage memory, eliminating the necessity for a garbage collector and ensuring consistent runtime performance. Through static analysis of variable lifetimes, the Mojo compiler efficiently frees data when it’s no longer in use.

Manual memory management

Additionally, Mojo presents a manual memory management system reminiscent of C++ and Rust, employing pointers for effective control over memory allocation and deallocation.

Matrix Multiplication in Mojo

In the Matrix multiplication example, we noticed a remarkable 17.5 times performance increase by importing Python code into Mojo.

Moreover, upon incorporating types into the Python implementation, the performance surged by an astonishing 1866.8 times.

Furthermore, employing techniques like Vectorizing, Parallelizing, Tiling, and Autotuning, they accomplished an incredible performance enhancement of 14050.5 times. This achievement is truly remarkable, surpassing even the optimization capabilities of languages like Julia and Rust.

Mojo Language Code Example

In this car example, we’ll craft a CAR class using Mojo syntax. First, we’ll establish a CAR class utilizing a struct and initialize variables with the ‘var’ keyword. To specify variable types, we’ll designate ‘speed’ as Float32 and ‘model’ as String. Since String is not a built-in type, we’ll need to import it.

Next, we’ll fashion two initialization functions, akin to Python but employing the ‘fn’ function. One function will solely accept car speed, while the other will accept both car speed and car model.

Following this, we’ll construct an object using CAR with a speed of 300. Finally, we’ll print out the car’s model.

Code Explanation and Output

– The code starts by importing the ‘String’ module.

– A structure ‘CAR’ is defined with two variables ‘speed’ and ‘model’.

– Two initialization functions ‘__init__’ are defined within the ‘CAR’ structure.

– The first ‘__init__’ function takes a single float argument ‘x’ and sets the ‘speed’ to ‘x’ and ‘model’ to ‘Base’.

– The second ‘__init__’ function takes two arguments, a float ‘r’ and a string ‘i’, and sets ‘speed’ to ‘r’ and ‘model’ to ‘i’.

– An instance of ‘CAR’ named ‘my_car’ is created with a speed of 300.

– The model of ‘my_car’ is printed, which should output ‘Base’ as the model was not specified during instantiation.

You can see that Mojo’s programming style and functionality resemble Python. It extends Python, improving its performance and memory management. Using Mojo, you can swiftly train models, achieve quicker model inference (even on CPUs), analyze vast datasets in seconds, and simulate in real-time.

To explore all the code examples, visit Modular Docs, and run them in a Jupyter notebook within the playground.


To sum it up, while Mojo looks promising as a fast language that can work alongside Python for AI and ML, it’s not likely to fully replace Python anytime soon. Python has a huge community and is deeply rooted in data science and ML. At best, Mojo could be an extra tool for situations where speed is super important.

In our tutorial, we went over Mojo’s features and how it works with some code examples. We also compared it to Python, looking at things like how fast it is, how its code looks, what it can do, and how well it works with other code. If you’re interested in learning the basics of AI and Machine Learning, we suggest enrolling in some Machine Learning Fundamentals with Python course.

What are perks of Mojo? Learn more

Leave a Comment