The RAM is soldered on all Halo Strix platforms because physics is getting in the way. With pluggable DIMMs the memory bandwidth would be halved, at best.
Without that, it's really not a interesting solution.
demanding replaceable ram means also not wanting the benefits of the integrated memory
Personally I think it's bad that apple products are so poorly repairable and so expensive to upgrade.
Soldered RAM, CPU, and GPU, that give space benefits and performance benefits is exactly what I want, and results in no more ewaste at all. In fact less ewaste, because if I had a smaller form factor I could justify keeping the older computer around for longer. The size of the thing is a bigger cause of waste for me than the ability to upgrade RAM.
Not everybody upgrades RAM, and those people deserve computers too. Framework's brand appears to be offering something that other suppliers are not, rather than expand ability. That's a much better brand and niche overall.
No. It's end of the line with consumerism and we either start repairing and recycling or we die. Framework catered to people who agree with that, and this product is not in line.
I have no idea why you would not upgrade your memory, I have done so in all PCs I ever owned and all laptops, and it's a very common (and cheap) upgrade. It reduces waste because people can then use their system longer, which means less garbage over the lifetime of a person. And as was already commented, it is not only about upgrades, but also about repairs. Ram breaks rather often.
Upgrading the RAM would have created more waste than properly sizing the RAM to COU proportion from the beginning.
It is very odd to encounter someone who has such a narrow view of computing that they cannot imagine someone not upgrading their RAM.
I have not once, literally not once have RAM break either. I have been part of the management of clusters of hundreds of compute nodes, that would occasionally each have their failures, but not once was RAM the cause of failure. I'm fairly shocked to hear that anybody's RAM has failed, honestly, unless it's been overlocked or something else.
Uncalled for and means the end of the discussion after this reaction. Ofc I can imagine that, it's just usually a dumb decision.
That you did not have to upgrade the ram means one of two things: You either had completely linear workloads, so unlike me did not switch to a compiled programming language or experimented with local LLMs etc. Or you bought a lot of ram in the beginning, so 8 years ago with a hefty premium.
Changes nothing about the fundamental disagreement with the existence of such machines. Especially from a company that knows better. I do not expect ethical behaviour from a bottom of the barrel company like Apple, but it was completely reasonable to expect better from framework.
One of the primary objections to soldered RAM was/is the cost to purchase. As the likes of Apple priced Ram upgrade at a hefty premium to retail prices.
But are Framework's RAM prices unreasonable? $400 for 64GB more of LPDDR5x seems OK. I haven't seen anybody object to Framework's RAM on those grounds.
My current laptop (ASUS GA503QM) had 8GB soldered and 8GB socketed. I didn’t want to go for the 16+16 model because it was way more expensive due to shifting from a decent GPU to a top-of-the-line GPU, and a more-expensive-but-barely-faster CPU. (I would have preferred it with no dedicated GPU—it would have been a couple of hundred USD cheaper, a little lighter, probably more reliable, less fiddly, and have supported two external displays under Linux (which I’ve never managed, even with nvidia drivers); but at the time no one was selling a high-DPI laptop with a Ryzen 5800H or similar without dedicated graphics.) So after some time I got a 32GB stick and now I have 40GB of RAM. And I gave my sister the 8GB stick to replace a 4GB stick in her laptop, and that improved it significantly for her.
You might be surprised. Living in a large city, everything I have put for sale has found a new owner. Old and seemingly useless computer hardware, HDMI cables that don't support 4K, worn-out cutlery, hairdryer that's missing parts, non-functional amplifier, the list goes on. If the price is right (=very low), someone has always showed up in person to carry these away. And I'm always very upfront about any deficiencies so that they know what they're getting.
I'd say a common profile for the new owner is young people who have just moved and are on a shoestring budget.
L2 CPU cache used to be on the motherboard and user expandable.
Would desoldering the sockets help?
Why are the sockets bad?
This sort-of-interview of Nirav Patel (ceo of framework) explains in a bit more detail: https://www.youtube.com/watch?v=-lErGZZgUbY
Basically, they need to use LPDDR5X memory, which isn't available in socketed form, because of signal integrity reasons.
Which means you won't see an improvement if you solder your ram directly, I think mostly because your home soldering job will suffer signal integrity issues, but also because your RAM isn't LPCAMM and isn't spread across a 256 bit bus.
Like physics PhD-level more.
This is also the reasoning why you can't just have a dumb female to female HDMI coupling and expect video to work. All of such devices are active and read the stream on the input and relay them on the output.
Above certain frequencies, you start running into major issues with signal integrity, and fixing them is very difficult without any extra circuitry.
I was commenting on a brand based on repairability selling a product that's deliberately not repairable. It's a curious choice to throw away the branding that brought them to where they are, and hopefully not the start of a trend for their other devices.
With an increased number of channels, you could have a greater amount of RAM at a lower frequency but at the same bandwidth. So you would at least be able to run some of these much larger AI models.
The only way to achieve what you're after is to do any of;
- Give up on unified memory and switch to a traditional platform (which there are thousands of alternatives for)
- Cripple the GPU for games and some productivity software by raising latency beyond the norm.
- Change to a server-class chip for 5x the price.
This is an amazing chip giving server-class specs in a cheap mobile platform, that fill a special nieche in the market for for both productivity and local AI at a very competitive price. What you're arguing for makes no sense.
For example Nvidia seek to ban consumer GPU use in datacenters as they to sell datacentre GPUs.
If they made consumer platforms that can take 1tb of ram etc, then people may choose to not buy EYPC.
Afterall many cloud providers already offer Ryzen VPS's.
We need a bigger memory controller.
To get more traces to the memory controller We need more pins on the CPU.
Now need a bigger CPU package to accommodate the pins.
Now we need a motherboard with more traces, which requires more layers, which requires a more expensive motherboard.
We need a bigger motherboard to accommodate the 6 or 8 dimm sockets.
The additional traces, longer traces, more layers on the motherboard, and related makes the signalling harder, likely needs ECC or even registered ECC.
We need a more expensive CPU, more expensive motherboard, more power, more cooling, and a larger system. Congratulations you've reinvented threadripper (4 channel), siena (6 channel), Threadripper pro (8 channel), or epyc (12 channel). All larger, more expensive more than 2x the power, and is likely to be in a $5-$15k workstation/server not a $2k framework desktop the size of a liter of milk or so.
This is the real story not the conspiracy-tinged market segmentation one. Which is silly because at levels where high-end consumer/enthusiast Ryzen (say, 9950 X3D) and lowest-end Threadripper/EPYC (most likely a previous-gen chip) just happen to truly overlap in performance, the former will generally cost you more!
However Apple will let you upgrade to the pro (double the bandwidth), max (4x the bandwidth), and ultra (8x the bandwidth). The m4 max is still efficient, gives decent battery life in a thin light laptop. Even the ultra is pretty quiet/cool even in a tiny mac studio MUCH smaller than any thread ripper pro build I've seen.
Does mystify me that x86 has a hard time matching even a mac mini pro on bandwidth, let alone the models with 2x or 4x the memory bandwidth.
(I suppose that you could devise a platform with support for mixing both "fast" in-package and "slow" DIMM-socketed memory, which could become interesting for all sorts of high-end RAM-hungry workloads, not just AI. No idea how that would impact the overall tradeoffs though, might just be infeasible.
...Also if persistent memory (phase-change or MRAM) can solve the well-known endurance issues with flash, maybe that ultimately becomes the preferred substrate for "slow" bulk RAM? Not sure about that either.)
The market dynamics are pretty clear. Having that much memory bandwidth only makes sense if you're going to provide an integrated GPU that can use that bandwidth; CPU-based laptop/desktop workloads that bandwidth-hungry are too rare. The PC market has long been relying on discrete GPUs for any high-performance GPU configuration, and the GPU market leader is the one that doesn't make x86 CPUs.
Intel's consumer CPU product line is a confusing mess, but at the silicon level it comes down to one or two designs for laptops (a low-power and a mid-power design) that are both adequately served by a 128-bit memory bus, and one or two desktop designs with only a token iGPU. The rest of the complexity comes from binning on clock speeds and core counts, and sometimes putting the desktop CPU in a BGA package for high-power laptops.
For Intel to make a part following the Strix Halo and Apple strategy, Intel would need to add a third major category of consumer CPU silicon, using far more than twice the total die size of any of their existing consumer CPUs, to go after a niche that's pretty small and very hard for Intel to break into given the poor quality of their current GPU IP. Intel doesn't have the cash to burn pursuing something like this.
It's a bit surprising AMD actually went for it, but they were in a better position than Intel to make a part like Strix Halo from both a CPU and GPU IP perspective. But they still ended up not including their latest GPU architecture, and only went for a 256-bit bus rather than 512-bit.