Show HN: We made our own inference engine for Apple Silicon

We wrote our inference engine on Rust, it is faster than llama cpp in all of the use cases. Your feedback is very welcomed. Written from scratch with idea that you can add support of any kernel and platform.

github.com

・

177 points

・

darkolorin

・

3 days ago

45 comments

TheMagicHorsey ・ 3 days ago

Amazing!

How was your experience using Rust on this project? I'm considering a project in an adjacent space and I'm trying to decide between Rust, C, and Zig. Rust seems a bit burdensome with its complexity compared to C and Zig. Reminds me of C++ in its complexity (although not as bad). I find it difficult to walk through and understand a complicated Rust repository. I don't have that problem with C and Zig for the most part.

But I'm wondering if I just need to invest more time in Rust. How was your learning curve with the language?

  • adastra22 ・ 3 days ago

    You are confusing familiarity with intrinsic complexity. I have 20 years experience with C/C++ before switching to rust a few years ago. After the initial hurdle, it is way easier and very simple to follow.

giancarlostoro ・ 2 days ago

Hoping the author can answer, I'm still learning about how this all works. My understanding is that inference is "using the model" so to speak. How is this faster than established inference engines specifically on Mac? Are models generic enough that if you build e.g. an inference engine focused on AMD GPUs or even Intel GPUs, would they achieve reasonable performance? I always assumed because Nvidia is king of AI that you had to suck it up, or is it just that most inference engines being used are married to Nvidia?

I would love to understand how universal these models can become.

  • darkolorin ・ 2 days ago

    Basically “faster” means better performance e.g. tokens/s without loosing quality (benchmarks scores for models). So when we say faster we provide more tokens per second than llama cpp. That means we effectively utilize hardware API available (for example we wrote our own kernels) to perform better.

zackangelo ・ 2 days ago

We also wrote our inference engine in rust for mixlayer, happy to answer any questions from those trying to do the same.

Looks like this uses ndarray and mpsgraph (which I did not know about!), we opted to use candle instead.

nodesocket ・ 2 days ago

I just spun up a AWS EC2 g6.xlarge instance to do some llm work. The GPU is NVIDIA L4 24GB and costs $0.8048/per hour. Starting to think about switching to an Apple mac2-m2.metal instance for $0.878/ per hour. Big question is the Mac instance only has 24GB of unified memory.

  • khurs ・ 2 days ago

    Unified memory doesn't compare to a Nvidia GPU, the latter is much better.

    Just depends on what performance level you need.

smpanaro ・ 3 days ago

In practice, how often do the models use the ANE? It sounds like you are optimizing for speed which in my experience always favors GPU.

  • AlekseiSavin ・ 3 days ago

    You're right, modern edge devices are powerful enough to run small models, so the real bottleneck for a forward pass is usually memory bandwidth, which defines the upper theoretical limit for inference speed. Right now, we've figured out how to run computations in a granular way on specific processing units, but we expect the real benefits to come later when we add support for VLMs and advanced speculative decoding, where you process more than one token at a time

    • J_Shelby_J ・ 3 days ago

      VLMs = very large models?

      • mmorse1217 ・ 3 days ago

        Probably vision language models.

ewuhic ・ 3 days ago

>faster than llama cpp in all of the use cases

What's your deliberate, well-thought roadmap for achieving adoption similar to llama cpp?

  • pants2 ・ 3 days ago

    Probably getting acquired by Apple :)

  • khurs ・ 2 days ago

    Ollama is the leader isn't it?

    Brew stats (downloads last 30 days)

    Ollama - 28,232 Lama.cpp - 7,826

    • DiabloD3 ・ 2 days ago

      Ollama isn't an inference engine, its a GUI slapped onto a perpetually out-of-date vendored copy of Llama.cpp underneath.

      So, if you're trying to actually count LLama.cpp downloads, you'd combine those two. Also, I imagine most users on OSX aren't using Homebrew, they're getting it directly from the GH releases, so you'd also have to count those.

      • imtringued ・ 2 days ago

        Actually, ollama has stopped using llama.cpp and is using ggml directly nowadays.

greggh ・ 3 days ago

"trymirai", every time I hear the word Mirai I think of the large IOT DDoS botnet. Maybe it's just me though.

  • fnord77 ・ 2 days ago

    I think of the goofy Toyota fuel cell car. I think a grand total of about 6 have been sold (leased) in california

zdw ・ 3 days ago

How does this bench compared to MLX?

  • jasonjmcghee ・ 3 days ago

    I use MLX in lmstudio and it doesn't have whatever issues llama cpp is showing here.

    Qwen3-0.6B at 5 t/s doesn't make any sense. Something is clearly wrong for that specific model.

rnxrx ・ 3 days ago

I'm curious about why the performance gains mentioned were so substantial for Qwen vs Llama?

  • AlekseiSavin ・ 3 days ago

    it looks like llama.cpp has some performance issues with bf16

sharifulin ・ 3 days ago

Wow! Sounds super interesting

skybrian ・ 3 days ago

What are the units on the benchmark results? I’m guessing higher is better?

dcreater ・ 3 days ago

Somewhat faster on small models. Requires new format.

Not sure what the goal is for this project? Not seeing how this presents adequate benefits to get adopted by the community

  • worldsavior ・ 3 days ago

    It's utilizing Apple ANE and probably other optimization tools provided by Apple's framework. Not sure if llama.cpp uses them, but if they're not then the benchmark on GitHub says it all.

  • koakuma-chan ・ 3 days ago

    Written in Rust is a big one for me.

slavasmirnov ・ 3 days ago

that’s exactly we are looking for not to waste on apis. Wonder how significant trade offs are

cwlcwlcwlingg ・ 3 days ago

Wondering why use Rust other than C++

  • adastra22 ・ 3 days ago

    Why use C++?

    • khurs ・ 2 days ago

      So C++ users don't need to learn something new.

  • khurs ・ 2 days ago

    The recommendation from the security agencies is to prefer Rust over C++ as less risk of exploits.

    Checked and Lama.cpp used C++ (obviously) and Llama uses Go.

  • outworlder ・ 2 days ago

    Why use C++ for greenfield projects?

  • bee_rider ・ 3 days ago

    I wonder why they didn’t use Fortran.

  • giancarlostoro ・ 2 days ago

    ...or D? or Go? or Java? C#? Zig? etc they chose what they were most comfortable with. Rust is fine, it's not for everyone clearly, but those who use it produce high quality software, I would argue similar with Go, without all the unnecessary mental overhead of C or C++