There continues to be a lot of incorrect information echoing around the Expansion Bay connector.
We’re developing our own semi-custom connector with the supplier specifically to make it better for end-user handling. The datasheet that people are referencing that states 50 cycles is for an off the shelf connector that Framework Laptop 16 doesn’t ship with.
I’m very much looking forward to the option that the Framework team can introduce to high-end AMD mobile GPUs." The 7700S only has 8GB of VRAM, which is definitely not enough to run games at native 2K resolution. Although I am a little disgusted with NVIDIA, I think the launch of the mobile version of the Quadro GPU can attract a professional users. The above is just a personal opinion.
You will have FSR2 and perhaps a 7800/7900s or whatever in the future.
This is mentioned in another thread, but I agree 8gb of VRAM is simply not enough - especially considering many will actually receive the laptop in 2024, and that it has a 1440p screen. It is fine to start with mid-range GPU, but there will be serious interest for more very soon.
The crucial issue will be price policy. If framework can deliver high-end parts at prices that take into account the large amount paid at the start, that would be fantastic. If the RTX 4080 module costs 990 Euros in Q3 of 2024, well, then the product is turned ad absurdum.
Incredibly complex issue I imagine, as probably finding partners is difficult, let alone the technical parts. But the success story of this laptop is still not sure, it will literally depend on the GPU.
can you make statement on a datasheet on how many cycles it does have and then link it everywhere?
There is still a pending deep dive blog post on the expansion bay. Maybe they’ll disclose this information there.
Yep. If they had hard test data he’d come out and share it already to finish that off.
Probably still engineering, picking one of several solutions, testing to complete etc… and our pre-order finalise emails are months away.
They may not have testing completed yet, but they do have good reason to believe they can do more than the 50 cycles people are still repeating.
Now, I must admit that I contributed in the thread where that number started. But it was only a listed as a minimum cycles number. I went ahead and deleted one of my posts where I repeated it while neglecting to mention that it was a minimum cycles number. Editing that post was not an option due to the post age, or else I would have struck it out & linked to Nrp’s post about it. And of course, I talk about the cycles being limited right in this thread, though without a number. Since an expansion bay deep drive is coming, we should probably just refrain from going on about it for a little bit, myself included. Apologies for what I contributed in repeating that 50 cycle number.
Has there been testing to suggest how many cycles we can expect the framework to support then? If it’s high, I could imagine some people will simply remove the screws and then hot-swap it anyways.
He says they are developing a semi-custom connector. Not that the process is completed with all testing 100% finished. And remember there is an expansion bay deep dive planned. We all need patience.
I doubt that would work. Probably need at least the two connector screws to compress the contacts. But who knows what we’ll see in the planned deep dive.
Also, the term “hot-swap” means doing it without shutting down the computer. PCIe would not cooperate well with that.
Correct. This is a compression-type connector, so it doesn’t function unless compressed using the fasteners.
Unfortunately I agree ;). I think the hopeful message of needing RTX 4080 like horsepower soon and in an affordable manner has been received long ago, and I think now it is up to the deep dive.
The actual GPU costs 450$. So i think if there would be a 4080 mobile, it would cost you way more and im not sure how “affordable” that would be (comparer to desktop)
If you dont need it mobile, you could use an eGPU and watch out for Oculink-Threads here in the forum.
Obviously Asus applies their own tax on their products, but look at the price of their RTX 4090 external GPU. Dread to think how much a similar module for this laptop would cost.
I read a lot of requests on this forum. I think people get the wrong idea about the FW16 and want to make it a desktop replacement. Its not that If you need maximum power for your workflow. A laptop is most likely not the solution for you.
Yeah, you pretty much have to ignore most of the request posts. So many, many people forget that Framework’s laptops have to be made to appeal to most people, not each user’s personal exact wishlist, including sometimes very niche use cases. This, by necessity, involves compromises. If Framework catered to every user desire, it would be a monstrosity. Would have 3 fold-out screens, weigh 45 lbs, and need a shoulder strap. And also 5 people max would ever buy it. 
Feature request: Make it also fit in my pocket. It’s not that hard, just do it! And don’t charge me for it!
I’m sure top-tier GPU modules are in the future though, so I think that’s a reasonable request. I can’t imagine myself ever needing that kind of power in a laptop, so thankfully they started with something more “humble” that will hopefully allow the battery to remain a relevant component.
Actually, I thought pretty hard about getting no GPU module and waiting for the first upgrade to come out so I can get a used 7700S for half the price, but I didn’t wanna wait. Ok, yes, Starfield may have also been part of the equation. But seriously, from what people are saying about the 780m performance, I might be able to live with just that, at least for a good while.
And there you have part of the problem though
I opted for no GPU module, not because I’m not excited about the possibility or have no use for it, but because the current version is not quite strong enough to justify it. The 780M is powerful enough for flatscreen gaming at low-mid with FSR, you’d only need the 7700S for mid-high gaming (probably also FSR with that DPI of the internal screen). So it’s not useless, but if you want a LOT of power, it’s still not it. I’d only want that for VR, for flatscreen a 780M is plenty IMO, so instead I’m now looking towards the possibility of an eGPU with Oculink.
For reference, some estimate the speedup of the 7700s over the iGPU to be 2.2-2.4x, given that it is a lower power desktop 7600 chip. In the grand scheme of things, 450€ extra on a 2000€ laptop for over twice the gaming performance is pretty good, on the other hand, a desktop 7600 is MUCH cheaper.
A very interesting take!
I went for it for the purposes of video editing, where that VRAM really helps out. If I were actually concerned with gaming on this, you definitely would have just talked me out of the module since personally, I’m not going to enjoy playing a game on a 16" screen much more just cuz it’s a higher FPS. My eyes suck and the size of the monitor is more important than anything for me when gaming, so if I’m playing something on a laptop I’ve either already accepted it’s not going to be fun and I just need the fix until I can get home, or it’s a game that doesn’t really need to look that great to be fun. In both cases I don’t need the performance boost of the 7700S anyway so if it’s not a make or break feature like it can be in video editing, I personally would not bother with it.
But I know there are lots of people who do rely on their laptop to play games, and I can still see the argument that the boost is totally worth it. Like you say, twice as many frames isn’t nothing!
Honestly, I feel it does slot into the mobile workstation segment fairly well… it compares to the HP ZBook G10 A at close to the same price point.