Computer RAM gets biggest upgrade in 25 years but it may be too little, too late — LPCAMM2 won’t stop Apple, Intel and AMD from integrating memory directly on the CPU::LPCAMM2 is a revolution in RAM, but it faces an uphill struggle
On CPU is definitely superior for performance, and what I’m not seeing people consider here is a future where you have On-CPU-RAM and On-Motherboard-RAM. CPU RAM for intense CPU functions, and traditionally seated RAM to be more like a modern “swap” I suppose, but instead of using the slower disks for swap, you’re just using slower RAM.
I could especially see this in Enterprise level hardware. I’m just saying, don’t throw the baby out with the bathwater. Por Que No Los Dos?
I know, I know, you can’t expect corporations to do squat to benefit the consumer, but one can hope.
Yeah, there is no way they’re gonna put 1TB of RAM on a CPU die anytime soon.
Does that mean that consumer hardware will include expandable RAM though? I feel like for the average person, that option still has a very high chance of disappearing on a lot of machines.
Oh yeah, a very high chance of disappearing. The unfortunate reality is probably 80% of people never upgrade their laptops or desktops. Building and maintaining your own PC has become more en vogue in recent years, but the vast majority of average consumers just don’t take part in the practice. Thus, it will not be prioritized by the industry. Why spend money on making your machines upgrade-able if the majority of users don’t ever take advantage of the feature?
I don’t like why it will happen, but I understand the economics of it.
I think most people don’t know the difference between “on-die” and “on-package”. This may be what they mean: https://beebom.com/intel-meteor-lake-cpu-on-chip-ram/
Also, lots of users aren’t gonna want the main system memory on the CPU die. Aside from the fact that it creates a clear path for vendors to artificially inflate prices through pretended scarcity via product segmentation and bundles, it also prevents the end users from upgrading the machines.
I’m pretty sure this even goes against the stated goals of the EU in terms of reduction of e-waste.
I have no doubt that a handful of vendors cooperating could restrict their offer and force the hand of end users, but I don’t think this would be here to stay. Unless it provides such a drastic performance boost (like 2x or more) that it could be enough of an incentive to convince the masses.
People who really care about computers buy handmade artisanal transistors.
I know you’re kidding, but: http://www.homebrewcpu.com/
wasnt there a guy who made his own GPU too?
Here is an alternative Piped link(s):
https://www.piped.video/watch?v=-vHwZhWoWkk
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
We already have memory wafers glued to our CPU wafers in the form of L3 cache. It’s lower latency, higher throughput, up to a few hundred MiB in bigger models and can potentially be used without external RAM sticks (but I’ve not heard of using that feature outside of BIOS firmware early boot – that’s probably the only change we’ll see). Sometimes it’s DRAM, sometimes it’s SRAM, its size varies quite a bit.
Wouldn’t RAM on die mean lower wafer yield?
This is about RAM on the package not RAM on the die. It honestly makes no sense why we don’t have CPUs and RAM soldered to the motherboard right next to the CPU package. I love being able to change the stuff myself, but any reasonable repair shop could be doing that for you and we can have much higher performance than we currently have. It’s not like there’s really many viable options anyways. AMD has what four good CPUs intel has like two, and there’s two good ram ICS.
Why would you think soldering would increase performance vs socketed at all much less provide “much higher performance”
If soldered was the only option ans 6 skud was enough for everyone everyone would have to buy very expensive hardware to increase one spec instead of smart people getting to mix match and upgrade.
The socket has big reductions in ram and pcie signal integrity. If you don’t plan to change the CPU and motherboard separately soldering it would save money and the store could do it for you when you order both together.
Because it’s true. Soldering the memory right next to the CPU allows us to run the memory at a lower voltage and faster clock rate, while getting lower latency too. The LPDDR4/5X are designed based around these improvements. GPUs have been doing this forever too for the same reasons. It’s a huge upgrade in every way except upgradability, which is effectively eliminated.
I’d like to see non-synthetic benchmarks showing real world performance increase in otherwise as close as possible to identical systems
Can it not just be socketed next to the CPU?
We seem to socket CPUs just fine.
On an ATX motherboard maybe, but there would be no space on a laptop or ITX board.
Question: modern systems can mount hundreds of GB or even terabytes of RAM, right? At this point, why not mount non-volatile storage as RAM? Performance should increase since data wouldn’t have to be loaded.
What you’re describing is the holy grail of computer memory technology. If we had nonvolatile memory as fast as RAM, we would absolutely be using it instead. Unfortunately even the fastest SSD today would be a significant drop in speed from modern RAM.
Would it be faster than loading gigabytes of data from an SSD over NVMe into RAM?
RAM is basically the scratch space for the CPU. It doesn’t just contain data loaded from the SSD, it contains the running state of the OS and all applications and is constantly being read from and written to by the CPU. As it is, RAM is already a lot slower than the CPU. Replacing RAM with a standard NVMe SSD would slow a PC down to an unusable crawl.
I’m familiar with what RAM is for.
The idea I’m getting at is, back in the day computer software especially video games came on ROM cartridges, which were little more than a plastic shell around a small circuit board that had a ROM chip or two, maybe a battery backed RAM chip to save game progress if you’re Nintendo. This ROM was attached to the same data bus on the CPU as the system’s built-in RAM, which meant the CPU could read from it just like RAM. This meant that cartridge-based systems didn’t need as much total RAM, because instructions, graphics, audio etc. were read directly from ROM, and it didn’t have to load instructions or graphics into memory before executing/displaying them. So no loading screens.
Modern computer SSDs are attached via SATA or at best the PCIe bus, and the CPU has to interact with a controller on the device to get data loaded into memory before it can be executed. Could you attach some non-volatile storage on the same bus with the RAM. Like, the average ATX motherboard has 4 RAM slots, right? Could you mount two sticks of RAM and two sticks of non-volatile storage that exists in the memory address space?
The simple answer your entire query is the fact that PCIe is not the bottleneck. The bus that connects your storage to the CPU is faster than any storage we have these days (ignore sata it doesn’t count). There is no innovation to be had with your suggestion, just complexity.
Now, there is something related to your suggestion. Graphics cards are exploring / have explored having storage communicate with them directly because of the latency in loading assets into the CPU then sending them to the GPU. This is where progress is being made on optimizing communication between storage and host.
deleted by creator
If you’re only loading data to access it once, then yes, but it almost never is the case - some specific programs might do it, but OS definitely caches pretty much everything it can in RAM for subsequent access - Linux, for example, fills unused RAM with cache
From the perspective of a computer engineer SSDs are painfully slow. Waiting for data on disk is slow enough that it is typically done by asking the OS for the data and having the OS schedule another process onto the CPU while it waits. RAM is also slow although not nearly as slow. Ideally you want your data in the L1 cache which is fast enough to minimally stall the CPU. The L2 and L3 caches are slower but larger and more likely to have the data you want. If the caches are empty and you have to read RAM your CPU will either do a lot of speculative execution or more likely stall.
Speculative execution on CPUs is a desperate attempt to deal with the fact that all memory access is slow by just continuing through the code as if you know what is in memory. If the speculative execution is wrong a lot of work gets thrown out (hopefully nothing unsound happens) and the delay is more noticable.
Bluntly an SSD only system would probably be an order of magnitude slower. I’m also not sure switching to a new process (or even thread) to load from SSD would be viable without RAM as it would likely invalidate a lot of cache triggering more loads.
Essentially we do. If you run out of RAM, you get pages from disk. You would know this if you ever used Windows ME.
3d vcache would like to say hi
both techniques will obviously need to coexist for some time. they dont have logistics on large memory near the processor,. quite yet, so there is still a place for ram.
deleted by creator
This is the future of PC, but with soldered RAM: https://en.m.wikipedia.org/wiki/Slot_1
