Technology
This is the official technology community of Lemmy.ml for all news related to creation and use of technology, and to facilitate civil, meaningful discussion around it.
Ask in DM before posting product reviews or ads. All such posts otherwise are subject to removal.
Rules:
1: All Lemmy rules apply
2: Do not post low effort posts
3: NEVER post naziped*gore stuff
4: Always post article URLs or their archived version URLs as sources, NOT screenshots. Help the blind users.
5: personal rants of Big Tech CEOs like Elon Musk are unwelcome (does not include posts about their companies affecting wide range of people)
6: no advertisement posts unless verified as legitimate and non-exploitative/non-consumerist
7: crypto related posts, unless essential, are disallowed
view the rest of the comments
Upgradable RAM isn't as fast as non-upgradable RAM and that this is especially true for the way Apple Silicon is designed. So no we shouldn't be mandating something that reduces computer performance for the sake of an upgrade most people would never care to perform.
We should however force them to produce laptops with a certain minimum RAM and to reduce their ridiculous upgrade pricing.
Edit: also I don't own a single Apple product. I aren't a fan boy at all and I know they do a whole bunch of anti-consumer bs. I also know that modular RAM for Apple Silicon would be a terrible idea for that specific design. Modular SSDs on the other hand would be very doable.
A quick look at the claims suggest 100GB/s is the RAM speed for the M2 Macbooks.
A single DDR5 RAM stick is about 50GB/s. So that's two of those in a dual channel config (effectively quad channel since each DDR5 stick is now a dual channel on it's own).
There's a good argument for introducing a new smaller DDR5 module so size isn't an issue, but I'm not sold on speed being the main problem. RAM is fast even when it's slow, and having more of it is almost always better than having it faster. No amount of RAM speed will ever compensate for swapping to storage when you run out.
At the very least mandate that the manufacturer replace the RAM at a reasonable cost at a later date, if you need more for future apps or if it goes wrong. We go on and on about fighting eWaste, yet entire laptops go in the bin when they don't have enough RAM.
Go look at the RAM speed of the M2 Pro and M2 Max. They are essentially quad and eight channels respectively to get the speed they achieve. Good look doing that with SODIMM modules.
Actually good RAM speed is absolutely essential for GPU performance. Saying how more RAM speed isn't important for a use case like the Apple Silicon Macs is ignorant AF.
You're getting heavily downvoted by people who obviously don't understand how RAM works. Or how computers work?
Guys, Apple is shitty, we all know this, but onboard RAM is the least of their anti-consumer practices.
The problem with socketed RAM is the length of the traces going back to the CPU. That 100% reduces performance (and battery life) by a significant amount. Especially when using that socketed RAM as iGPU VRAM.
Dell's CAMM standard reduces the latency compared to SODIMM, for socketed RAM, but what we really need is for someone like Apple to invest R&D into really tiny RAM sockets that are super close to the CPU, instead of researching ways to lock users out.
Doesn't even sound that complex. Little LGA style socket, tiny heatsink clip to hold it in place.
There's even laptops that have soldered RAM and a SODIMM slot. Could you limit the GPU to using the soldered RAM? Still won't help you if it develops a fault though.
The RAM is built onto the substrate. Every contact you add increases signal degredation. Plus actually trying to fit eight sockets on a SoC package would be a complete nightmare.
Dividing RAM like that into two pools would violate the permise of the whole unified memory system. You're really asking for the wrong thing here. Why not convinve them to do something like a modular SSD that's far more achievable? Also memory that doesn't come at sky high prices with an actual sensible mimimum (8GB on MacBooks in 2023, really?).
For other laptops there is actually a solution to this problem called a CAMM. It would even work for the M2 Macbooks possibly (not the M2 Pro or Max) if apple are willing to sacrifice size or battery life of the laptop. The reason this wouldn't work for the M2 Pro and Max is you would need two or four of these things. It would be diffcult enough to fit just one in a Macbook that have tiny, tiny logic boards to begin with.
Thanks a bunch. The level of ignorance here to Apple's design choices is palpable. Some of the stuff they do is very anti-consumer. Soldered RAM isn't one of them - at least on Apple Silicon. Having modular GPU RAM hasn't been a thing for over a decade for good reasons.
I doubt the difference in performance is that significant. If it was 50% faster then sure. But odds are it's something like 3% speed difference. Same for the storage, I doubt that apple's proprietary interface is that much faster than a regular high quality nvme, definitely not enough to justify the multiple that they're charging for it compared to an off-the-shelf nvme.
Erm yeah it's more than 50% faster in bandwidth for M2 Max, because it has more memory channels than two SODIMMs would allow for. It's specifically at least twice as fast. People upvoting this are showing their ignorance here about Apple hardware.
The storage isn't particularly fast so that part I believe.
Really? Why though? Is soldered-in RAM attached differently to the CPU?
Way differently.
Soldered RAM is much much closer to the CPU, and so the time it takes for signals to propagate back and forth is significantly reduced..
It's probably the increased capacitance (think of of it as a puddle that needs filling before the water can move beyond it) of a mechanical connnection system vs direct soldering that makes most of the difference.
I was going to call you out on the distance thing but I made the maths and indeed at 100GHz light only travels about 3mm between waves and electric signal propagation on a line is roughly lightspeed (if you disregard capacitance) so even though this memory bus is likely not working at 100GHz to get 100GB/s (it's actually using paralellism for increase bits per cycle) it is none the less already within clock speed ranges were distances of centimeters do mater.
That said keep in mind that rountrip propagation only really maters at the very biginning of the download of a memory block as that when the address goes down and the data starts coming back and the roundtrip propagation affects the delay between them.
But yeah, I can see how you would start worrying with centimeter and even millimiter distances when trying to extract a bit more performance from data exchanges at these clock speeds.
This is an argument that just gets repeted. My question is this, is a macbook faster than a gaming pc? Because that has replaceble ram, cpu, gpu, ssd, etc. If yes, then please seek help.
The PC GPU does have it's own soldered RAM. But then the performance of a good GPU goes way past that of a MacBook, which while good for integrated graphics, is still only on par with a GTX 1660, a four year old budget GPU.
Well fucking said dude. You know dGPUs used go have upgradable RAM? They removed it because it dosen't work for that application. Apples iGPUs struggle to compete even being soldered partly because the competition is using GDDR and they aren't. Not soldering would make them even further behind.
Erm yeah. Have you never seen an M1 chip? It's on the same substrate.