Well, i loved the concept. Love the laptop, guess i’ll just have to hope it lasts a long time.
Because of the following;
-Decided to launch an entirely NEW modular laptop, framework 12 rather than new framework 16 parts. Further dividing resources and labor from a small outfit.
-Will be selling a desktop that is an antithesis to their upgradable reusable mission. More or less a paperweight once the l.l.m. bubble marketed as, sigh, A.I. bursts. The only redeeming factor of the machine is up to 128gb or ram. Despite being soldered on and ram does wear out. so maybe a igp with a ton of ram?
-Absolutely no word on the framework 16. Suspicious considering they even had linus tech tips talk about it before release.
-And no announcement of a new gpu module after the announcement of new gpus from ‘both’ amd recently. And Intel at the end of last year.
Sadly it seems the project is either shelved or abandoned for whatever reason. Sad to see.
I too felt that way when they didnt announce anything about the Framework 16.
But I have to admit, that cmonkeys answer on reddit sums it up quite nicely:
There won’t be updates on things that aren’t ready yet.
But no updates dont mean, they arent working on anything.
And while my anxiety doesnt like that, it also gives me hope that we can just “let them cook”
Additionally there isn’t a cpu with the correct power levels (including ones with configurable TDPs) and enough IO to support a dgpu, 2 ssd’s and all the USB IO
We know they are working on the 16 or they wouldn’t be shopping around with different vendors for new keyboard options. There are also 240W USB-C chargers out on the market now and I would be surprised if Framework wasn’t working on launching their own as soon as they can find a good/reliable manufacturing partner.
Right now is just a weird time for the 16 as a category because of what it is.
For the Framework 16 to get full function out of all of the expansion cards, 2 nvme slots, and the interposer for the expansion bay it needs the full 20 PCIE lanes from the 7840HS/7940HS. Everything announced with AMD’s Ryzen AI line only has 16 PCIE lanes and would require some degradation of functionality. Intel’s new Core Ultra line would make up for it with things like their H skews falling within TDP budget and having a whopping 28 PCIE lanes with native Thunderbolt support, but we also don’t know if Framework agreed to any exclusivity deals with AMD with how involved they were with 16, even being part of the AMD partner program, but that is just speculation.
As for an upgrade to the 7700S we have strikes from all three players. AMD and Intel haven’t announced any new mobile GPUs outside of their integrated solutions attached to their processors and the general scuttlebutt on the forums is that nVidia is controlling with their partners and would be unlikely to work with Framework on the upgradeable approach. Maybe they could try a 7800M or 7900M when they release the 240W charger but I don’t know how happy people would be with getting “old silicon” for more than the 7700S already costs.
We might just be looking at largely sitting with what we have for a while.
That’s exactly the problem.
There are so many small problems since months and no fixes yet.
What about newer drivers? BIOS updates? Thermals?
We don’t even talk about new parts.
*looks at all the pre-release publicity. Blogs, emails, features on LTT despite being an ‘investor’ such things cost over 100k+ per video. This went on for over a year with multiple videos from MULTIPLE creators well before the product launch.
*looks at all the information people are claiming ‘they’re doing’ but no official word has been released. Nothing, just crickets. To me that’s more wishful thinking, and possibly sunk cost thinking. I too paid over 2k for it and that’s why i’m disappointed.
Along with.
*mobile gpu’s are not big events anymore and aren’t even announced till laptops actually ship with them. Haven’t been since they became just better voltage-gated versions of the lower end desktop gpus. Which means the silicon is binned by how well they work at lower voltages, thus less power. Just like you can buy UL(ultra low voltage, vender undervolted) versions of some intel chips
Less pcie lanes, yes, but still ‘newer’ mobile chips. Better performance. And above all. ->AN OPTION<- That could be presented for people who ‘want’ it and to help start off the ecosystem.
How many of the fw16 used the second m2 slot AND the gpu? The 13 has similar trade-offs depending on what cpu you choose. Looking at you risc-v.
Don’t get me wrong, I don’t know if a single new anything will ever be released by Framework for the 16 laptop. If not, then Framework is simply lying when they say they have a long road map of continued support for it. But sure, it’s possible they’re just lying.
But the fact is, they almost never reveal plans and upgrade road maps ahead of time. I saw talk exactly like this back when all they had was the FW 13, about the fact that they didn’t offer an AMD processor. People talked about how they didn’t offer an AMD processor and despite that people had been saying they wanted an AMD option for two years, Framework released a new Intel mainboard generation, but still hadn’t said anything about AMD, so obviously they just were never going to offer an AMD processor for various, speculated reasons. And then, the AMD 13 laptops and mainboards were announced for pre-order.
So sure, maybe they’ll never release anything else for the FW 16. But the fact that they haven’t announced anything yet certainly doesn’t mean they won’t.
I think it is quite difficult for FW.
The FW16 uses 17x PCIe lanes. 8x for the dGPU and 2x 4x for the M.2 slots, and 1x for Wifi.
It has some other lanes, but those are used for USB 4 / Thunderbolt.
Unfortunately, the latest mobile chips from AMD only have 16x pcie lanes.
The 7840HS currently used in the FW16 has 20 pcie lanes.
So, until there are mobile chips that support more than 16x PCIe lanes, its going to be difficult to upgrade the FW16.
16x PCIe lanes is enough for probably 99% of the laptops out there. Like the FW13 and FW12.
Let me restate. I am sure people would be ‘fine’ with one m2 slot instead of two. Boom, you can now have a newer cpu on a newer arch and keep everything else.
OR
Maybe someone has no inclination to have a d-gpu, but still likes the larger size of the 16, They can drop the dgpu and use those lanes for the two m2 slots. on mb’s for those people who want a newer cpu, because they use the laptop for cpu things rather than something that require a non igp.
OR
They can use these. PI3DBS16413 PCIe Gen 4 Mux/DeMux Switch - Diodes Inc | Mouser
Pcie multiplexer chips exist. Doing what they did in late 90’s with irq’s. Splitting lanes like they used to have virtual irq’s for systems where you have more devices than free irqs.(god i remember the nightmares BEFORE that…)
The AMD fire range CPU has enough PCI lanes for everything the current framework 16 laptop has and more: up to 28 PCI lanes including some PCI Gen 5 for the ssd. These processors use the suffix HX and HX3D with up to 16 cores.
The “problem” is that they are rumored to come with an iGPU with only 2 CU… so they more or less require a dGPU.
The first laptops using it are supposed to be announced in April… they are “workstation” laptops and will be pared with RTX 5090. Examples here Laptops
As you can see they are not cheap.
…So if that’s what framework plans to use for the framework 16, they might announce it soon. We will have to see what dGPU framework decides to use…
Framework’s desktop PC (which is advertised on the Framework laptop under a category entitled, ‘Framework Laptop’ - sic) has soldered RAM? Goodness, it does. Surely that is the ‘antithesis’ of Framework’s (own!) philosophy. True, this is a small desktop, but soldered = unrepairable and indeed unreplaceable (unless one replaces the whole mainboard, I take it).
The x86 mobile processor space has been pretty boring for the past 18 months or so (especially in the sub-30w category), IMO. Both Intel and AMD are trying to get their houses in order to come up with something that’s fast, and energy efficient. M4’s single core score is around 50% higher than Intel’s mobile processors. AMD’s single core is even slower. Both camps eats watts like there’s no tomorrow. And there’s also the mix of iGPU and NPU performance in the storm. This isn’t really an Apple vs Intel vs AMD. It’s just there’s nothing to call home for while they try to come up with something new.
Hell, it’s almost like a battle between RISC and CISC chip design.
The RX7900M is an PCIE Gen4x16 Card and needs 180w (160-200w TGP) alone.
The Graphicsinterposer is PCIE Gen4x8.
The Power is doable, according to Frameworks Paperwork + FW 16 Connector Deep Dive the Interface is capable of 10A at 20V (200w) and its unlikely that FW will release a GPU which is not using its full PCIE Lanes. There is a RX7800m floating around, but i don’t think thats an viable Option either. I hope for an RDNA4 Release as soon as the Mobile Chips arrive. Navi44 maybe
re the graphic.