Produced by | OSC Open Source Community (ID: oschina2013)AI programming language Mojo🔥 has launched a version for the Mac platform, with its founder Chris Lattner stating that Mojo + Apple Silicon is a powerful combination.
Mojo🔥 is a new programming language developed by Modular AI, combining the ease of use of Python with the portability and performance of C, aiming to make it an ideal choice for AI research and production.
Download link: https://developer.modular.com/downloadDevelopers need to register for a Modular account first, then download the Modular CLI via the Homebrew package manager, and run the command <span>module install mojo</span> to install Mojo.
Below is a screenshot of running “Hello Mojo”:
Additionally, the Mojo SDK also provides a Visual Studio Code extension.
The Mojo team states that the Mojo language can fully utilize Mac CPU cores and vector units for acceleration. According to their tests—running a matrix multiplication example using <span>matmul.mojo</span>—on the Apple MacBook Pro M2 Max, Mojo is approximately 90,000 times faster than the pure Python implementation.
Open source developer Aydyn Tairov ported <span>llama2.py</span> to Mojo last month—<span>llama2.mojo</span>. At that time, only the Linux version was available, and the results were surprising—he stated that Mojo’s SIMD primitives helped improve Python’s poor performance by nearly 250 times.Now that Mojo for Mac has been released,Aydyn Tairov claims<span>llama2.mojo</span> performs on par with <span>llama2.cpp</span> on Mac, and in many cases even outperforms plain C.
Related links:
-
https://www.modular.com/blog/mojo-is-now-available-on-mac
-
https://twitter.com/tairov/status/1714103695321829551
-
https://twitter.com/clattner_llvm/status/1715038783832526874
Previous recommendations
Quiz | New paradigms of machine learning in the era of big data The US government restricts the export of NVIDIA H800 GPUs to China A master crafts a RISC-V simulator with Scratch, successfully running the Linux kernelHere are the latest open source news, software updates, and technical insights
Click here ↓↓↓ Remember to follow✔ and star⭐