- cross-posted to:
- hackernews@lemmy.smeargle.fans
- hackernews@derp.foo
- cross-posted to:
- hackernews@lemmy.smeargle.fans
- hackernews@derp.foo
Apple has a memory problem and we’re all paying for it::Apple still sells expensive “Pro” computers with just 8GB of RAM and charges a fortune for more.
Definitely not true hardware-wise. L2 cache is different from DDR3 RAM is different from DDR4 RAM… in price and performance
Software-wise, yes, the operating system abstracts away the differences and memory is memory
Apple’s memory upgrade costs are probably 90% usual Apple bullshit pricing, 10% grounded in reality. I’m thinking that the 10% may be something like the motherboards are designed without memory upgrades in mind, so if you want more RAM, they have to use a special mobo which they prefab less of
Apple’s “memory upgrade” is making the claim that you can do with half for the same amount of work on x86. It is 100% untrue.
You tried to delve into speed. But speed won’t outpace a 1TB video file you’re trying to edit. If you’re working with smaller chunks of smaller files that have fast operations ONLY, then make claim as such. This is a ploy for upgrade cash, plain and simple. Nothing about these chips moves the needle on memory usage BY HALF. What a dumbass thing to assert as a company.
Apple’s claim isnt 100% untrue tbh. It depends on the operations actually. Arm processors have at least 12 registers to contain data or references to memory. A program does need more ram space on a x86 processor, as it only has like 4 registers afaik ( correct me if im wrong! ) so it needs to push data more often to the stack.
This means that the m* processors has to generally save less info in memory. However, data is still data and you still need memory to contain the data youre processing so you still need the ram. So like, when doing video work that apple claim is utter bullshit. Raw calculations however might not be so much bullshit
For many memory intensive operation, this is incorrect since by that logic, Apple’s chip should use far more memory due to having quarter as many registers for those purpose. (32x64 vs. 32x256)
Most processors have cache memories for reasons you stated.
Correct, cache exists for that reason. But youre still loosing time and space by saving it to memory ( cache is just faster acccess for the cpu. Its still in ram or in the pipeline to be pushed to ram on next flush ).
Also true, per thread you would need more memory to save the cpu’s state when switching threads. Now i kinda want to do some calculations and tests to see at what point it gets better.
I always figured that per thread more memory is needed, but that the thread itself needed less memory ( or time to access it ) because it can contain more temp values in the cpu’s registers.
Again though, there claim is bullshit or not totally depends on the kind of work youre doing and for video work i consider it bullshit as well :')
A semi isn’t a Ford F150. That’s the end of that argument.
Memory speed doesn’t really matter if your apps start thrashing
Edit: thrashing is very likely to occur on something marketed as “pro”. I have a work PC with 8gb of RAM, and my job requires me to edit video. I need to be careful on how big my video files are, because it WILL start thrashing. This is the reality. Professional apps require a lot of memory pages, and they are never open on their own.
Edit 2: I guess the thoughts from a computer scientist are less important than corporate marketing.
Apple uses a unified memory where the memory chips are embedded on the SoC in the first place. The memory modules are on the same silicon wafer the chip is cut from, not separately on the Mobo, and shared directly with the chip in a single pool of memory that the CPU and GPU can access, rather than dedicated memory for each.
Changing the memory means cutting a different piece of silicon for it.
The SoC and memory are separate dies with different manufacturing processes. In the case of M2 it was TSMC for the SoC and SK Hynix for the memory.
When it comes time to package them together, the SoC and memory are soldered to a interposer layer. So the only difference is which size memory chips they solder together for the different memory configurations available.
That’s like building a fast car that can only go straight. It’s impressive but short-sighted and therefore stupid.
There are specific performance benefits to soldering your memory to the board or making them part of the main die itself. It’s why GPUs have been doing it for a long time, and why laptops with soldered ram can often achieve higher clocks and lower latency than their socketed counterparts. It’s a tradeoff, but a calculated one. I’m sure Apple also adds the extra revenue from absurd upgrade costs into their calculations.
Somebody has never been to a top fuel drag race. Impressive does not even begin to describe what those “fast cars that can only go straight” are capable of.
Ha ha. Most people don’t use their laptops exclusively for one single thing. I sometimes need a laptop that can go fast but more often it needs to be able to many different things. And some years later, let me swap in some more RAM and an SSD to give it another few good years.
This is 100% false. All Apple Silicon Macs use standard LPDDR4X or LPDDR5 memory chips, the same as are used in other computers, which are soldered on a PCB next to the SoC. They are not on the same die. The high memory bandwidth on M1/M2/M3 comes from having a lot of memory controllers built into the SoC – it’s akin to a PC with an 8+ channel memory setup. As far as I’m aware, there’s nothing technically preventing Apple from making an Apple Silicon mac with socketed memory again, other than those sweet sweet profits for shareholders.
I mean why let them bullshit even 10% ?
“There might be some hard to find small benefit here, maybe.”
That just sounds like you want it to be true, but deep down you know it isn’t.
Shit like this is apple play book from the nineties, especially “we need less ram” and “our clock cycles are better”.
It wasn’t.