Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

MacBook Pro with M5 Pro and M5 Max

apple.com

Post reply on HN
Print view View original

51–60 of 1001 posts

Re: MacBook Pro with M5 Pro and M5 Max

#51

Post by carlosjobim »

testfrequency wrote: post #43
carlosjobim wrote: post #36

Earlier quoted context omitted.

You might have confused Hacker News with your e-mail inbox again. This is an Apple press release, directed to everybody in the world who might be interested in a new computer or their first computer.

What’s with the attitude? My machine is aging like a fine wine, I’m acknowledging how resilient their custom silicon is despite the world demanding more and more compute.

It was a joke, should have put a smiley face. But every thread on a new Apple product here on HN have the same "why should I upgrade" comment, forgetting that there are people who might have very old devices they want to upgrade, or they might want to switch from Windows/Android to Apple.

Even if a new device is a small upgrade from last year's model, it can be a giant upgrade for other people.

Re: MacBook Pro with M5 Pro and M5 Max

#52

Post by fny »

whizzter wrote: post #26
Tangokat wrote: post #5

"Scaling up performance from M5 and offering the same breakthrough GPU architecture with a Neural Accelerator in each core, M5 Pro and M5 Max deliver up to 4x faster LLM prompt processing than M4 Pro and M4 Max, and up to 8x AI image generation than M1 Pro and M1 Max." Are they doubling down on local LLMs then? I still think Apple has a huge opportunity in privacy first LLMs but so far I'm not seeing much execution.…

We had a workshop 6 months ago and while I've always been sceptical of OpenAI,etc's silly AGI/ASI claims, the investments have shown the way to a lot of new technology and has opened up a genie that won't be put back into the bottle. Now extrapolating in line with how Sun servers around year 2000 cost a fortune and can be emulated by a 5$ VPS today, Apple is seeing that they can maybe grab the local LLM workloads if…

I feel like the push by devs towards Metal compatibility has been 10x than AMD. I assume that's because the majority of us run MacBooks.

Re: MacBook Pro with M5 Pro and M5 Max

#53

Post by butILoveLife »

dirk94018 wrote: post #37

On M4 Max 128GB we're seeing ~100 tok/s generation on a 30B parameter model in our from scratch inference engine. Very curious what the "4x faster LLM prompt processing" translates to in practice. Smallish, local 30B-70B inference is genuinely usable territory for real dev workflows, not just demos. Will require staying plugged in though.

[flagged]

Re: MacBook Pro with M5 Pro and M5 Max

#54

Post by Aurornis »

testfrequency wrote: post #15

I have a fairly maxed out M2 Ultra (24 cores, 192GB RAM), and still cannot get this machine to choke on anything. I have not once felt the need to upgrade in years, and that’s with doing pretty demanding 3D and LLM work.

I have a powerful older Mac that doesn’t really “choke” on anything, but I could always use more speed.

The high memory Macs have been great for being able to run LLMs, but the prompt processing has always been on the slow side. The new AI acceleration in these should help with that.

There are also workloads like compiling code where I’ll take all the extra speed I can get. Every little bit of reduced cycle time helps me finish earlier in the day.

And then there’s gaming. I don’t game much, but the M1 and M2 era Apple Silicon feels sluggish relative to what I have on the nVidia side.

Re: MacBook Pro with M5 Pro and M5 Max

#55

Post by eknkc »

dirk94018 wrote: post #37

On M4 Max 128GB we're seeing ~100 tok/s generation on a 30B parameter model in our from scratch inference engine. Very curious what the "4x faster LLM prompt processing" translates to in practice. Smallish, local 30B-70B inference is genuinely usable territory for real dev workflows, not just demos. Will require staying plugged in though.

I find time to first token more important then tok/s generally as these models wait an ungodly amount of time before streaming results. It looks like the claims are true based on M5: https://www.macstories.net/stories/ipad-pro-m5-neural-benchm... so this might work great.

Re: MacBook Pro with M5 Pro and M5 Max

#56

Post by snowchaser »

whizzter wrote: post #6

128gb of memory, it's a nice change for Apple not to lag in that department for once, wonder what such a machine will cost though.

In US, going to 128 GB from 32 is $1500 extra. However 32 GB is only offered with the 32 core version and 128 only with the 40 core version.

Re: MacBook Pro with M5 Pro and M5 Max

#58

Post by replwoacause »

testfrequency wrote: post #15

I have a fairly maxed out M2 Ultra (24 cores, 192GB RAM), and still cannot get this machine to choke on anything. I have not once felt the need to upgrade in years, and that’s with doing pretty demanding 3D and LLM work.

Sounds pretty beefy. What kind of local LLM is that thing capable of running? Does it open up real alternatives to cloud providers like OpenAI and Claude, or are the local models this hardware is capable of running still pretty far behind?
Post reply on HN