261–270 of 1001 posts
I use a 48" OLED with my MacBook Air M3 and for me that is a near ultimate web development experience both on desktop and when travelling: https://bsky.app/profile/benhouston3d.bsky.social/post/3l7li...
I love the innovation Apple has brought with their investment in ARM. That said, I can't imagine buying a computer in the 21st century that can't be opened and upgraded, especially with a price premium attached. I just don't get it. I am in no way trying to be combative, but I'd love to hear a counterpoint that makes sense for these machines.
And as someone who works with computers all day long, I never ever want to open one up unless I build one from scratch and want a personal project.
Earlier quoted context omitted.
Sure, but to be honest, I don’t think you should be shipping a high-spec computer with a built-in screen. Anything high-spec should be broken into at least a separate screen and computer. The screen is such a major point of failure. I’ve seen so many iMacs and MacBooks with broken screens where the end solution is to replace the entire device, which is a waste. It’s that much more of a waste if you are getting a high…
Yeah, that's my biggest complaint with iMacs: I want to be 100% certain I can repurpose them as monitors later in their lives. I have a ten year old iMac at the moment that would make an amazing second monitor... but it doesn't quite have the features I need to use it like that. An HDMI input would be great.
I’m contemplating doing this myself at some point but my maxed out 2019 iMac upgraded to 128GB ram and extra SSD is still plenty fast for me, actually feels subjectively quicker than my M2 Pro MacBook Pro with significantly less ram feel. I was a bit surprised as I had read all the hype of the responsiveness of the Apple M-machines.
Earlier quoted context omitted.
Sure, but to be honest, I don’t think you should be shipping a high-spec computer with a built-in screen. Anything high-spec should be broken into at least a separate screen and computer. The screen is such a major point of failure. I’ve seen so many iMacs and MacBooks with broken screens where the end solution is to replace the entire device, which is a waste. It’s that much more of a waste if you are getting a high…
> The screen is such a major point of failure. The screen of an iMac (at least the 5K 27" one) is the part that still has value, so its annoying when you can't use it on new hardware.
Earlier quoted context omitted.
We should do a "Most HackerNews HackerNews Comment Of The Year" award each year; in homage to the O.G. [1]. I'd like to nominate this comment. [1] https://news.ycombinator.com/item?id=9224
I agree with the premise but this isn't a great nomination, I think saying "this device won't appeal to the niche LLM gadgetry crowd" is valid in this case since yeah, it won't.
As someone who mucks around with running LLMs and other large models on my computer the 32GB maximum RAM is a show-stopper for me. I'm on a M2 with 64GB at the moment and I'm already regretting not going for 96 or even 128. I want to be able to run a large model AND other apps at the same time.
I really wish that "elite" LLM folks like yourself or many others in the field would just abandon apple/mac. Yes, they make better laptops from a hardware perspective. No, M series chips are not actually competitive with Nvidia hardware on anything except cheap (and relatively slow) inference of big models. The fact that windows laptops are considered DOA despite the insane amount of inertia behind Nvidia GPUs is jus…
Until then it’s meaningless to even consider windows (laptop)
As someone who mucks around with running LLMs and other large models on my computer the 32GB maximum RAM is a show-stopper for me. I'm on a M2 with 64GB at the moment and I'm already regretting not going for 96 or even 128. I want to be able to run a large model AND other apps at the same time.
We should do a "Most HackerNews HackerNews Comment Of The Year" award each year; in homage to the O.G. [1]. I'd like to nominate this comment. [1] https://news.ycombinator.com/item?id=9224
https://news.ycombinator.com/item?id=35079
> pg: That has to be the comeback of all time.
Do they sell matching monitors to go with these all in one style iMac's? I'm not really an Apple person, but I can't imagine having a single monitor in this day and age and any other color would look silly next to a weird pastel pink device.
No. If you’re that kind of power user, the iMac isn’t for you. Get a Mac mini or studio.
Earlier quoted context omitted.
The ram is on the package for more than portability. It’s necessary for fast enough transfer speeds for the iGPU.
> necessary for fast enough transfer speeds Source?
LPCAMM solutions mentioned upthread resolve some of this by making the problem more "three dimensional" from what I can tell. They reduce the length of the traces by making the pinout more "square" (as opposed to thin and rectangular) and stacking them closer to the actual dies they connect to. This allows you to cram swappable memory into the same form factor, while retaining the same clock speeds/size/bus width, and without as many design complexities that come from complex socket traces.
In Apple's case they connect their GPU to the same pool of memory that their CPU uses. This is a key piece of the puzzle for their design, because even if the CPU doesn't need 200GB/s of bandwidth, GPUs are a very different story. If you want them to do work, you have to feed them with something, so you need lots of memory bandwidth to do that. Note that Samsung's LPCAMM solutions are only 128-bits wide and reported around 120GB/s. Apple's gone as high as 1024-bit busses with hundreds of GB/s of bandwidth; the M1 Max was released years ago and does 400GB/s. LPCAMM is still useful and a good improvement over the status quo, of course, but I don't think you're even going to see 256-bit or 512-bit versions just so soon.
And if your problem can be parallelized, then the higher your bus width, the lower your clock speeds can go, so you can get lower power while retaining the same level of performance. This same dynamic is how an A100 (1024-bit bus) can smoke a 3090 (384-bit) despite a far lower clock speed and power usage.
There is no magical secret or magical trick. You will always get better performance, less noise, at lower power by directly integrating these components together. It's a matter of if it makes sense given the rest of your design decisions -- like whether your GPU shares the memory pool or not.
There are alternative memory solutions like IBM using serial interfaces for disaggregating RAM and driving the clock speeds higher in the Power10 series, allowing you to kind of "socket-ify" GDDR. But these are mostly unobtainium and nobody is doing them in consumer stuff.
I guess it is just me, but I cannot stand working on small screens on the desktop. I will accept a small screen on a laptop because it is inconvenient carrying around something larger, but on a desktop, there is no reason to compromise. I use a 48" OLED with my MacBook Air M3 and for me that is a near ultimate web development experience both on desktop and when travelling: https://bsky.app/profile/benhouston3d.bsky.s…