Hacker News Reader

Comment by olliej | Hacker News Reader

Wait, so are they using the GPUs or the "neural engine"? because the gpu approach should also work on any other machine right?

View on HN

The Apple ML Research group announcement has more details: https://machinelearning.apple.com/updates/ml-compute-trainin...

Key line:

  Until now, TensorFlow has only utilized the CPU for training on Mac. The new tensorflow_macos fork of TensorFlow 2.4 leverages ML Compute to enable machine learning libraries to take full advantage of not only the CPU, but also the GPU in both M1- and Intel-powered Macs for dramatically faster training performance.

So, looks like it's faster on both Intel & M1, but the M1 MBP has a much faster GPU than the Intel MBP

But note that the ANE and the GPU are different facilities.

Correct - I’d guess this is Not using the neural engine.

I don’t know enough about that hardware to hazard a guess about how easy it would be to get that part of the chip involved.

Looks like Tensor Flow now supports ML Compute, which in turn uses the appropriate hardware.

In the case of the Mac Pro, I'm guessing ML Compute is using the GPU.

In the case of the Intel MacBook Pro 13", which as far as I can tell from Apple's site can't be purchased with a discrete GPU, that will be either the Intel Iris GPU, or the CPU.

In the case of the M1 MacBook Pro 13", I'm assuming ML Compute prioritizes the Neural Engine over the GPU (and CPU), but don't know if there are use cases where the GPU would be preferable.

It looks like the class only has cpu and gpu types so, as has been the case, figuring out when the 'neural engine' is invoked is remains opaque:

https://developer.apple.com/documentation/mlcompute/mlcdevic...

As the charts compare Intel based Macs vs the new M1 chip, they would be comparing CPU performance.

The charts are comparing “accelerated” performance versus what appears to be a cpu only baseline. It is not clear which M1 hardware beyond the CPU is used

Reply on news.ycombinator.com