681–690 of 1001 posts
Still super interesting architecture with accelerators in each GPU core _and_ a dedicated neural engine. Any links to software documentation for how to leverage both together, or when to leverage one vs the other?
Earlier quoted context omitted.
Serious questions. How is Asahi these days? Is it ready as a daily driver? Is it getting support from Apple or are they hostile to it? Are there missing features? And can I run KDE on it?
nevermind. Found this. Still a ways to go. https://asahilinux.org/docs/platform/feature-support/m4/#tab...
Earlier quoted context omitted.
And the fastest M4 max was already fastest single and multicore CPU by a decent margin, while the fastest non-Apple CPU was only specialized for single or multi.
The single thread performance for modern high performance CPUs are all very close to each other. Apple's latest usually has a small advantage because they're the first to use TSMC's latest nodes, which is good for something like 15-20%. The fastest multicore CPUs are the ones with a lot of cores, e.g. 64+ core Threadrippers. These have approximately the same single-core performance as everything else from the same ge…
Earlier quoted context omitted.
Yeah I love my M1 iPad Pro. But the "liquid glass" update made it feel slower. Really only the 'unlock' feels slower, once I'm using it it's fine. But it's slightly annoying and does make me want to update this year to the m5. But it's a glorified Kindle and YouTube box, so I'm hesitating a little bit.
my dad's got a pre AS iPad Pro and it's so bad after updating to 26. My 6th gen iPad on iOS 17 felt faster than this
Which is fine, since it’s exclusively used to watch a kids show for a half an hour a day.
…but it’s also super sad to see a once fantastic piece of kit to degrade so much primarily due to software.
It’s unfortunate that this announcement is still unspecific about what they improved in the Neural Engine. Since all we know about the Neural Engine comes from Apple papers or reverse engineering efforts ( https://github.com/hollance/neural-engine ), it’s plausible that they addressed some quirks to enable better transformer performance. They have written quite interesting papers on transformers on the Neural Engine:…
That's the only way to speed up MLX 4x compared to M4.
Earlier quoted context omitted.
Serious questions. How is Asahi these days? Is it ready as a daily driver? Is it getting support from Apple or are they hostile to it? Are there missing features? And can I run KDE on it?
nevermind. Found this. Still a ways to go. https://asahilinux.org/docs/platform/feature-support/m4/#tab...
Base models only: - M1 | 5 nm | 8 (4P+4E) | GPU 7–8 | 16-core Neural | Memory Bandwidth: 68.25 GB/s | Unified Memory: 16 GB | Geekbench6 ~2346 / 8346 - M2 | 5 nm (G2) | 8 (4P+4E) | GPU 8–10 | 16-core Neural | Memory Bandwidth: 100 GB/s | Unified Memory: 24 GB | Geekbench6 ~2586 / 9672 - M3 | 3 nm (first-gen) | 8 (4P+4E) | GPU 8–10 | 16-core Neural | Memory Bandwidth: 100 GB/s | Unified Memory: 24 GB | Geekbench6 ~296…
Interesting to see that over 5 years (M1 was 2020), the benchmark performance has not quite doubled. Is this an indictment of Moore's law, or just Apple over-speccing the M1 and slowly decreasing that over time?
Earlier quoted context omitted.
I’m trying to find any reason I can that my M1 Max needs replacement; it’s hard. How do you justify it?
You should wait until next Fall if you don't really need to replace your M1 Max. Rumors say that Apple's going to redesign the Macbook Pros next year with an OLED screen.
A lot of Apple hardware is impressive on paper, but I will never buy a Mac that can't run Linux. I simply don't want to live in Apple's walled garden. Then there is the whole ARM vs x86 issue. Even if a compatible Linux distro were made, I expect to run all kinds of software on my desktop rig including games, and ARM is still a dead end for that. For laptops, it's probably a sensible choice now, but we're still far f…
> A lot of Apple hardware is impressive on paper, but I will never buy a Mac that can't run Linux. They run Linux actually very well, have you ever tried Parallels or VMware Fusion? Especially Parallels ships with good softwaer drivers for 2d/3d/video acceleration, suspend, and integration into the host OS. If that is not your thing, the new native container solution in Tahoe can run container from dockerhub and co.…
Windows - because I needed it for a single application.
Linux - has been extremely useful as a compliment to small arm SBCs that I run. eg: Compiling a kernel is much faster there than on (say) a Raspberry Pi. Also, USB device sharing makes working with vfat/ext4 filesystems on small memory cards a breeze.