← Back to context

Comment by aurareturn

1 day ago

Don't buy the Mini or Studio. Both have the M4 which lacks the Neural Accelerators, making prompt processing ~3-4x slower.

4 comments

aurareturn

Reply
mortenjorck  1 day ago

I assume those don't just work automatically with an off-the-shelf gguf. What do you need in your local inference stack to take advantage of M5's neural accelerators?

  • wren6991  14 hours ago

    Apple muddied the waters by calling them "neural accelerators" but it seems like what they actually added in the M5 generation is tensor instructions for the existing GPU cores. It's not a separate accelerator like the ANE.

    llama.cpp's Metal backend does use them when they're available.