Every time Framework is brought up, there's a chorus of people saying they won't buy until they release an AMD edition, and as one of them this is great to see.
Edit: Also the input deck and the ability to attach an egpu as well as upgrade to a newer egpu in the future are huge. This is a fantastic leap forward for Framework.
Yeah the possibilities are exciting to say the least, the ability to specialize your laptop into an AI processing powerhouse on the fly could be very valuable.
I’m one of the chorus. I actually preordered a TOTL Asus Zephyrus Duo about a month ago, and it came in last week. It’s been a dream so far — 16 real cores of AMD 7945HX, laptop 4090, hybrid graphics, two real screens! Main display is a quite color-accurate 16:10 miniled with 240hz freesync, lower display is a high-PPI IPS touch panel, dual raid-able M2 slots (I put in dual WD SN850Xs in raid0 and that’s showing ~14.5GB/s reads and ~13Gb/s writes in CrystalDiskMark), decent sound and webcam, keyboard is at the bottom edge…the only downsides are the touchpad size is a bit strange (portrait orientation) and the click on it doesn’t feel quite Macbook nice, no USB4, and the power brick is very brick-like, with a thick cable that doesn’t flex very well. Cooling is excellent due to the intake fans under the second screen combined with the lower heat output of the AMD chip, allowing it to run maxed out without throttling…they did this right. I can’t get Pop_OS to install yet, guessing it might need the AMD raid driver like Windows did, requires further investigation.
Anyway, excited for a Framework version too! While I prefer AMD integrated graphics to Intel, NVidia dGPU would be even better, and even better still if it was upgradable. Are laptop GPUs still available on MXM cards?
Why even try to use the hardware raid? I can't imagine it actually performs any better than mdraid, and with mdraid the drives are as portable as plain drives. You could destroy the special laptop and stick the drives into usb enclosures and access the array again on any other machine.
Tell me you work in ops with out saying you work in ops. :-)
Btw, this is excellent advice. Funny story, I have a FreeNAS device and the motherboard died and I thought "Oh my, I need to bring my ZFS volumes up on another machine, but I didn't have another machine with 6 SATA bays! I ended up having the drives all sitting out on the workbench connected a mainboard with 8 SATA ports so that I could create an archive of the data, and then got the mainboard fixed so I could re-assemble and re-use the FreeNAS but still it alerted me to the fact that I really needed a 6-8 drive cabinet if I wanted to do this again.
That's only until the machine in question is 5000km away and the soonest time you can get to it is in the three months.
Sure, for personal usage there is almost no usage for the HW RAID, but when you need to make sure what the system would always boot and it can't be serviced in hours/days - then you have almost no options for SW RAID.
No problem to put /boot on a raid1 on a small partition across all drives, so that any drive can boot, and no problem to even include a whole self-contained remotely accessible recovery os. It's a little bit more work to set up, but if you are professing a need for that, then a little extra setup is de rigueur. I remotely administered a ton of linux boxes in racks scattered across the US like that for years. Although I had out of band serial console access and could do full bare metal reinstall that way, I could also do it from any neighboring machine that was still running in the same rack if I had to, with a combination of network booting and/or booting from any one of the normal drives normal raid1 copies of the /boot partition.
Further remote-able fallback options that I never even had to use but could: Local hands just plucks a hot-swap drive from any of my other machines and pops it into the bad machine. All drives had the same bootable partition and all drives are redundant and so they could yank literally any one from the wall of server fronts. Or, better, local hands just plugs in a thumb drive and I take care of the rest. Thumb drive is already sitting there for that purpose, or they could make a new one from a download. But with 8 to 24 hot-swap drives per machine, meaning 8-24 copies of /boot, I never even once needed local hands to so much as plug in a thumb drive.
There is just no problem at all with sw raid. It only provides options, not remove them.
Did you even read my comment? It's quite clear what your environment was in the data-centers, with spares and remote hands.
Mine wasn't and then I say three months I don't kid or jest.
> No problem to put /boot on a raid1 on a small partition across all drives
This is exactly the problem. If the drive isn't totally dead (like it doesn't even respond to IDENT) then there is a chance what the BIOS/UEFI would try to boot from it and even succeed in that (ie would load the MBR/boot app) and then there is no way to alter the boot process at this point. HW RAID card provides a single boot point and handle the failing drives by itself, never exposing those shenanigans to the upper level.
Like sure, you are happy with your setup, you never had a bad experience with it, you always had OOB management and remote hands - but it doesn't means what it a silver bullet working 100% of times for everyone.
Yes, I've seen systems with SW/fake RAID failed to boot because the boot process failed after selecting a half-dead drive as a boot device, with my own eyes. Thankfully I was geographically close to them, not 5000km away.
Yes, I serviced and prepared systems for the 5000km away divisions and they are really serviced only a couple of months in a year, all other time you need an extremely urgent reasons why do you need to a rent a heavy 'copter to go there. No, there is no remote-hands there. The maximum point of IT-competency there is raking bills with satellite Internet.
Heh, you were "lucky": I had a 4-drive NAS that died on me, and no motherboard lying around at all, let alone one with 4+ SATA ports.
I bought four cheap SATA-USB3 adapters, plugged them into two USB3 hubs, which I then plugged into the two USB3 ports of a Raspberry Pi 4, and arranged it all quite precariously in a small cardboard box that I cut holes into for airflow. Performance was terrible, of course, but it worked well enough until I could build a proper new NAS box.
The portability doesn't end with sata or usb ports.
With generic software raid (mdadm), even if you only had a single usb port and a single internal drive, you could image all the drives one at a time and then access the array of images on the single big drive. (not uncommon since usually time has passed by the time something fails)
It also goes the other way if you needed it to. Say a single array member was 4T but you only have a bunch of 1T drives, no problem, you can assemble 4 1T drives into a 4T container to hold the 4T image, and then use that image as an aray member itself.
Even if you don't have any loose external drives, you could even do it all via network shares with pieces residing on all of your other family members laptops and desktops, or every desk in an office, while they all still continue running windows and doing their normal jobs I mean.
Some of the possibilities are slow or fragile or both, so of course you don't set out to use 12 usb2.0 ports, but the point is essentially anything is possible, and you don't have to worry about predicting or planning for every possibility, you just don't have to care about how you'll recover the array in the future because it doesn't matter what form storage takes at that time, or what form you happen to have available. It would almost never make sense to do some things, but the point is that mdadm just doesn't care.
For a machine with only 2 or 4 internal drives where you want to use raid0 for max throughput, and don't want to rely on any special firmware support for booting raid0, just partition the drives so that /boot is a small raid1 across all the same drives, so that any of the drives could boot. Bonus, it automatically makes all the members of your main raid0 slightly smaller than the drive's nominal size, which means you can always fit them onto some other replacement drive later, even if the different manufacturers count bytes and formatting overhead differently.
I come from the days of sco unix on scsi hardware raid with full featured expensive cards and I do not miss it.
Good call. I have manually repaired a few mdadm arrays in my time that would have surely been complete losses in the hardware RAID systems I've encountered.
+1.
Years ago I had a Raid1 mirror that failed because the controller itself went bonkers, and both drives were accumulating errors. Luckily I was using mdraid and could recover their files by using testdisk [0] on both of them separately as USB drives on a different Linux machine. Was a really really long process; I pray the storage divinities not to have to live those couple days again.
That is a pretty niche laptop, I think the Framework folks are going more for the every day driver kind of experience. I've got a Framework 13 and really like it, it is, for me, a pretty solid Thinkpad replacement. But the real icing on the cake will be when I update the motherboard for an upgrade. (since it is relatively new I don't expect that to happen until maybe next year).
On the Asus, are both screens touch screens? Or is only the lower screen a touch screen? I had looked at 20:2 type touch screens to do a sort of "media bar" setup on my desktop but didn't find anything at the time I could use. I'm wondering how well such a setup might work.
Really exciting to see. I almost got a Framework for my work setup last month, but lack of an AMD option held me back. My personal laptop is still chugging along, but the AMD Framework is at the top of my list now!
If only I'd waited an extra month to order my new work equipment.
There is always going to be a chorus about something. Power users are impossible to please because they all have strong opinions but not the same opinions.
People will still want different screen resolutions, aspect ratios, keyboard layouts, etc. Now that AMD is available people will want particular CPU SKUs. It will never end.
I wish Framework luck, because they are targeting the pickiest userbase that exists.
I personally held off on getting one until my previous device (a Spectre x360) got smashed up in a car crash, but don't in any way regret this purchase even if I do still sort of miss touchscreens (and probably also would have waited for an AMD option otherwise).
But yeah, when I was suddenly in the market for a new laptop anyway, the extensive freedom to upgrade and configure this made it an extremely easy sell compared to anything else that would have been on my radar.
AMD 3:2 thin-and-light at a small overall screen size (but still usable) and with replaceable parts checks an enormous amount of power-user boxes. Getting off Intel was their last critical hurdle, and while for whatever reason it took them an eternity to listen to that (rather overwhelmingly loud) feedback, they finally did it.
I suspect the choruses will be much quieter now. If I hadn't already landed on a Beelink GTR6 mini-PC desktop + MNT Reform laptop, I'd be considering this finally - indeed, one of many reasons I went Reform over Framework originally was because Framework was married to Intel. And frankly, 3:2 still makes Framework tempting despite my lack of need for it.
I think USB4/TB3 was a huge box for the latest gen AMD mobile platform. Can't speak for others, but I'm generally docked with power over TB3 at home, and when I was in the office before. It's the one thing I wanted out of AMD that was missing for general use. Glad to see it here.
Definitely going Framework for my next laptop in a couple years... really tempted to get one sooner than later and pass down my M1 air to my daughter.
They’re targeting that user base in exactly the right way: give me options, get out of my way, and let me do it.
Too many platforms lock users into decisions they think are best for them, or worse, that they think are best for the company making said platform. It’s refreshing to be given not only options, but options I actually want.
AMD have been thrashing Intel in performance, thermals and battery life for several years now. Intel are also a rather unpleasant company, I'll happily go with a competitor whenever it makes sense.
On Linux AMD drivers are by far some of the best, and the driver support is miles better both on the end of the 1st party support as well as 3rd party software support.
There's a few instances where AMD GPU drivers even outshine their windows counterparts like in OpenGL performance. Intel's driver support is great as well, but their performance recently has not matched AMD, and their iGPUs don't compare to AMD's iGPUs.
Also for beefier laptops with discreet GPUs, intel up until very recently would need an NVIDIA card to get decent performance which would require NVIDIA drivers, which are notorious for complicating things on Linux. On the other hand AMD laptops with discreet AMD gpus already have their drivers built into the kernel which removes almost any potential complication.
Also there's personal preference, which in my case leans toward AMD because of their generally more open stance with their new technology, like raytracing and FSR which is contrasted by NVIDIAs typically closed off approach to the same things. Intel isn't as bad as NVIDIA in this department either but this is just personal opinion.
Lastly AMD CPUs are generally cheaper than intel, and in the last few generations performed better overall than intel.
TL:DR AMD on linux has great drivers with good support, the synergy between AMD's CPUs and GPUs and their combined APUs are very good, and personally I appreciate AMD's more open approach to technology like FSR, even if they aren't anywhere near perfect.
When Ryzen first came out, Intel still used the number of cores to differentiate their consumer and enterprise lines, which meant you could get a bucketload of cores from AMD without paying "enterprise" prices.
AMD still tends to have lower actual TDP than Intel, which is more efficient in a laptop and easier to cool in a desktop.
If you're going to use the integrated GPU, RDNA generally performs better in games.
I have linux a notebook with the AMD Ryzen 7 5700u and a Mat screen. Its pretty great on battery (6+ hrs). Even the built in gpu is good enough for general use. I think these newer AMD cpus are even better and worth the wait.
My home linux machine is a ninth gen intel with Nvidia. Its technically a notebook, but more like a portable workstation. Its powerful, but battery life is terrible (<3) and you can hear the fans. It can game quite well however.
I wrote this a while back comparing Intel 12th Gen to AMD 6000, and power/perf should apply even more to the 13th Gen vs 7040 (Zen4 + RDNA3):
The reason that people want AMD CPUs is simple - they’re much better than what Intel is offering. We can see empirically how wide this gap is now since we have independent reviews of identical laptop platforms from Lenovo and HP’s ultrathin business laptops (the same segment as Framework):
HP EliteBook 840 G9 1280P vs HP EliteBook 845 G9 6950HS [1] - these are both respective flagship parts, and we can see the AMD version has both a +23% performance rating and a +23% better battery runtime. Note, that on the HP website currently, the same configuration AMD version is also >30% cheaper.
Lenovo ThinkPad T14s G3 6850U vs 1235U [2] - since the Intel version reviewed isn’t a top of the line model, it’s not fair to compare performance numbers, but even with the significantly lower TDP Intel part (15W vs 28W), the AMD version of the laptop ends up with an even bigger lead with a 35% better battery runtime. The AMD version is also 10% cheaper than their Intel counterparts from the Lenovo site pricing.
In both these cases, the AMD version wins significantly on processing power, battery efficiency and price. The 7040 should extend the lead on the latter, and that's not taking GPU performance into account [3], where the AMD Radeon 680M simply crushes the Intel Xe 96EU by an average of +88% in game performance, and +135% in synthetics (and again, the 7040’s RDNA3 GPU and Xilinx AI core will extend the lead even more for the upcoming generation).
Well, I have USB3 problems with all my AMD machines which makes e.g. attaching a VR headset reliably difficult. Random quickly repeated disconnects of devices due to a buggy chip from ASMedia/ASUS that AMD uses everywhere. I like TDP of the current gen though AMD turned to an even worse company than Intel in how it treats Threadripper users. DisplayPort over USB-C might not work either.
While having hardware compatibility issues suck, I'm not sure what your problems have to do with AMD's mobile APUs, which have their USB2/3/4 controllers implemented on-die (block diagram: https://cdn.videocardz.com/1/2021/12/AMD-Rembrandt-Diagram.j...) not using an external controller.
USB features depend ultimately on the laptop manufacturers' implementation (eg neither the HP or Lenovo models linked implement USB4 but they are in other models like the Lenovo Z13), but I haven't seen any that claim to have DP-alt support not be able to support it. As the Framework will have USB4 (and DP-alt and PCIe encapsulation are mandated by Microsoft for Windows 11 compatibility), I don't think that's a real worry for anyone.
Glad to hear that! My 8-core AMD laptop from ASUS with 3080 is almost 2 years old now so things might have changed. My Threadripper still has USB issues and I don't like e.g. inability to type text or move mouse for 10-20 seconds just because the USB controller on the most expensive Threadripper board decides to have a game of disconnects.
Again, I agree that hardware compatibility issues suck... Personally, I think life's too short to live with glitches like that - ASMedia chips are used in a lot of PCIe USB cards these days as well, but I've had good luck w/ Renasas cards in the recent past (when I was running VFIO). These USB cards are all pretty cheap ($20-40), so I'd order a few to try and keep the one that works the best.
I for one was waiting for a touch screen option as well, but this sounds easier than the AMD part sounded awhile ago. And they will probably add it as an option at some point, their track record for listening to what their users are saying in their forums is awesome so far.
I wonder what exact tech stack are they using for the website? It genuinely seems to me that this kind of a website should be able to handle this level of traffic on a single $10/month VPS, nevermind require multiple servers. But I'm probably just out of touch with how modern webdev is done.
Just put up a page with specs, no videos are needed. I have a framework laptop already which has been great, maybe one day there will be other laptops with no fan that are powerful like an apple arm.
This. Cloudflare is dead simple to set up, but curling your own site and setting up a redirect to a cached version of your pages is also trivial for anyone who can competently operate nginx / apache.
When the Framework came out, I was excited, but not in the market for a laptop. I've been waiting for the chance to buy one, especially as they announced more and more things that I'm interested in. Now that I'm finally in the market, I was waiting eagerly for this new announcement.
The only question I'm left with now, which Framework to get? Do I pre-order the AMD 7040, or do I wait for the new 16 so that I can get a GPU? I'm thinking about going the AMD route now, buying the dGPU option they have and using it as an eGPU when it becomes available, and then switching to the 16-inch chassis when I'm ready...
The AMD 7040 series is gonna be amazing. They even have an NPU which (theoretically) should run models reasonably well without a dGPU, M1 style... but we will see how that shapes up.
The 16" will provably perform better though, if you can wait.
The 780M integrated graphics on the 7040 apu is supposed to be something like 30% better than in the current-gen 680M. Search YouTube for 680M gaming and see if it's good enough for you!
It's going to be exciting to see which GPU manufacturer they went with. Intel salvaging arc? This eGPU that can be slotted in is exactly the thing I'd been hoping they would do if they added GPUs and the AMD option is incredible. Gonna have to put my money where my mouth is once it's time for a new laptop now...
I also think it's important to not underestimate the flexibility and marketing aspects of the eGPU design they went with. Someone who has a laptop they use for playing games with a still fine CPU but outdated GPU might look at this and just buy the eGPU. Then when it comes time to upgrade the laptop itself they might be all-in on the value of an upgradable system.
Yea, that’s true in a world where no other company is capable of manufacturing a GPU enclosure with the physical specs they’ve defined. Thankfully, I’m pretty skeptical we live in that world.
If the framework laptop grows in market size that there’s enough demand for it, then anyone could manufacture and sell those components, which is kind of the point.
We’ve opened the specs specifically to avoid this scenario! The information we have up on GitHub today is enough to get started on module development, and we’ll share any additional information that is useful to enable third party board makers.
My concern is more about business than technical factors. The NRE for such a module is going to be significant so the addressable market needs to be pretty big to justify it. Third-party modules could also get killed off by first-party ones that come later.
Third-party modules could also get killed off by first-party ones
I think it is pretty clear that Framework intends to allow you to buy the most barebones laptop you can. I mean they let you buy 0 of the little cartridge modules, I assume they won't force people to buy one of these PCIex8 modules.
I mean if some company is selling, say, a Radeon 7600 module and then Framework comes out with their own "official" 7600 module then the third-party company is toast.
Alternatively, that means Framework doesn't need to spend resources building their own version, or acquiring a company that makes one. It's to Framework's benefit that they don't compete with module vendors. They just need to let the vendor list them in the official marketplace and promote them as a success story for the modular laptop concept.
With this in mind, trampling module vendors doesn't seem like a great idea. What's the point of creating an ecosystem if there's nobody around to participate?
Some major issues with the livestream too, audio out of sync, horrible noise, looked like it was filmed with a potato (lighting was godawful). Did not exactly inspire confidence, but the upgrades do look nice. Glad there's an AMD option now.
Probably a good architecture for launch is a Cloudflare Pages static landing page that has the CTA link to your website that goes through a queue intermediary (invisible most of the time). That way you ensure you don't drop any visibility and you capture all sales.
Festival and concert tickets are frequently done this way and the queue renege rate lets you model how many you would have lost if your site failed to load and at what duration.
Programmable per-key remapping, layered modifiers that you can toggle per app or workflow (ie. quick language layout swaps, temporary low-level remaps for games that don't support remapping, Blender/Photoshop/editor-specific remaps for personal ergonomics, etc.). Tap-hold functions to send different keycodes or modifiers when you tap vs. hold down a key, often used for home-row mods that move modifier keys into the home row when you hold, say, an HJKL key.
The starfighter is a platform generation behind, and very expensive. $1600 would buy you a (same generation) Asus G15 with a huge dGPU and good linux support.
TBH I am kind of shocked Framework picked up the AMD 7040 series so quickly. Usually small manufacturers are stuck with older gens.
Starfighter doesnt have upgradeable / repairable RAM so it doesn’t compete in the same niche as Framework. Starfighter doesn’t look nearly as repairable.
Starfighter doesn’t have the proven “longevity” and true commitment to mission that Framework has had time to demonstrate, now through its second major release.
That all said, Starfighter looks like it has some really cool features between open source boot firmware, removeable camera and Wi-Fi kill switch.
Apples and oranges, kind of. Frameworks are repairable and have some hardware transparency, but aren't open, libre, free-as-in-freedom, or privacy focused like StarLabs. (EDIT: Looks like the 16" Frameworks are 16:10 like the Starfighter, vs. the 3:2 13" Frameworks.) Starfighters are Linux-first, Frameworks are Linux-compatible, and there are rougher edges on Linux Frameworks.
If you want a better or more libre Linux laptop right now, get a Starfighter. If you want a laptop with more hardware hackability and replacement options, at the expense of libre/privacy features, battery life, supported Linux options, and high-end performance, Framework's an option.
I hadn't heard about the ARM Thinkpad, that's very interesting. For me the reason ARM on the Framework is an exciting possibility is that you can switch out the mainboard with the processor. The X13s is $1,300 which is quite the investment into such an early platform. With a framework laptop, you could have an ARM mainboard, but with the ability to switch it out with an x86-64 mainboard without having to invest in a whole new laptop. It could also go the other way, having a laptop on hand and wanting to dive into ARM, you would just have to buy an ARM mainboard instead of an entire laptop.
ARM is great in theory, but I don't see any SoC on the horizon that I would really crave. This includes Apple, as I am either stuck with OSX or (at best) have most of the interesting bits of the SoC nonfunctional in linux or Windows.
Oh there we go, modular GPU add in boards. Heyoooo!
And they support other stuff!
I wish the framework site was up, as I want to see what this specification looks like. If you look inside an Asus Zephyrus or whatever, the cooling setup needed to support such a thing in a thin laptop is mad. The 13", for instance, has a pancake sized 2D headspreader + liquid metal TIM, which very much goes against repairability. The 16" is a wall of heatpipes. I wonder how framework is addressing this.
I repasted and changed the fans in my Zephyrus (turns out mattress pumps and fans don't mix) and fortunately the whole heatsink assembly comes off in more or less one go.
It is bulky though. Moreover, this bulk is not enough to keep the keyboard from getting really hot under load.
Quick summary for those couldn't watch the livestream/load the website.
- New Framework laptops are available using Intel 13th Gen and AMD 7040 series.
- New 61Wh battery, new matte display option & a Cooler Master mainboard case for $39.
- New 16 inch Framework laptop. Allows you to have a number pad or not, up to you and customizable. If I understood the livestream correctly, there are upgradable dedicated GPUs that connect to the 16 inch laptop.
The big question is whether the cooling system in the 13" would be able to handle the HS. Is there any data on Zen 4 laptop chips that has tested the efficiency vs performance considerations on the U vs the HS series?
On previous generations: the 4000 series starts to hit a relative performance wall at around 35W. The 5000 series clocked higher and hence gained more moving from 15 to 35 and 35 to 45.
Another point: you can throttle the 35W HS series to 15W, and get the laptop to run fanless or almost fanless since it is designed to cool a higher TDP.
Matte display yay! I missed that a lot on the MBP. Seems like they still want to bundle a windows license, I never understand this, do MS pay OEMs to do this?
More likely, the answer is that a lot of Windows users want a hassle-free experience. DIY edition of this would have no Windows tax, if you are a linux user.
If you pick the DIY option, you can omit the Windows license. You'll have to install Linux yourself, but that doesn't seem like a terribly difficult task for someone already going with the DIY option.
I'm guessing they don't want to officially support any specific distro, and offering support for multiple distros is probably a bit outside their wheelhouse at the moment. [e: this is perhaps not the case]
We continue to focus on solid Linux support, and we’re happy to share that Fedora 38 and Ubuntu 22.04 will work fantastically out of the box for both the 13th Gen Intel® Core™ and AMD Ryzen™ 7040 Series models. We have thorough setup and troubleshooting guides and will continue to provide official support for these two distributions. Manjaro XFCE 22.0 and Linux Mint 21.1 are also working great, and we’re detailing the documentation for those as well. You can check compatibility with popular distros as we continue to test them on our Linux page or in the Framework Community.
For what it's worth, Linux works great on the current generation. Even sleep works, although it wakes up so quickly that I'm not sure it's actually gone to sleep. I should try pinging it to see.
With the S0 suspend, it can actually go either way on if its' actually sleeping or low power mode... Not speaking to framework or any linux variant specifically, just know it's often an issue.
Sorry, I missed where it was specifically about the AMD model. The 13th-gen Intel got added to the list between me posting and you posting, so who knows.
Yes, several distros are supported now. Even NixOS, though not yet officially. As someone who purchased multiple DIY editions running linux, avoiding shipping any framework with a preinstalled linux distro has some real benefits:
Not shipping with Linux installed gets Framework out of the business of managing distros and images themselves. They also dodge a neverending barrage of "But what about my favorite distro X?" that would surely come for them as soon as they picked some fixed set of distros.
If you are uncomfortable assembling computer components and installing an OS yourself then linux is probably not for you. This acts as a filter for people who are serious.
It's a lot easier to support newbie computer users on windows, and they are a lot more likely to get support from friends and family.
They would be shipping a DIY edition anyways, this vastly simplifies their inventory.
These may offend your Linux-maximalist sensibilities, c'est la vie.
Some of us are perfectly comfortable with the software side, but no longer have the hand eye coordination or dexterity to handle those tiny little ribbon adapters in modern computers.
Personally, I kind of wish System76 would sell framework laptops with a generalized configuration and PopOS pre-insalled.
So how about 'No OS preinstalled' option? and allow customers to chose if they want to pay the MS tax or not. Maybe they are not allowed to offer this option?
Including any linux distro would be better, because it would show a growing linux desktop market share.
Depends how your Windows was licensed; if you had OEM license with license key that you had to type in (this is usually when Windows license is an option), or if the key was embedded in BIOS (when Windows license is provided and it is not an option).
Not OP, but in my case, I'd love an HS part without a dGPU. I'm typing this on a 5650U and its integrated GPU is overkill for what I do. Hell, until a year ago, I used to daily drive an i5-6500 with an HD530 with a 4k display and I never felt limited by the GPU.
But I do use the CPU from time to time, so a beefier part would be useful. I would rather not pay for a dGPU which I'd use maybe 1% of the time, and which would cost more and weigh more and possibly be a PITA to manage under Linux.
I dunno, as time goes on I am wanting to do more and more on the GPU. Last gen it was media processing and upscaling, now its running Stable Diffusion and other Generative AI.
But what it sounds like you want is an E-core mad part from Intel. I always thought they should sell a 20-30 core laptop part with only one or two big cores as a compilation monster.
> But what it sounds like you want is an E-core mad part from Intel. I always thought they should sell a 20-30 core laptop part with only one or two big cores as a compilation monster.
Indeed. I've been looking at the 12th gen laptops we're getting at work, but those were underwhelming. I don't remember which model CPU it was, but it only had 2 P cores (those are basically for "office work", so no one cares about performance). For compilation (Rust) it was in the same ballpark as my 11th gen i7 (so 4 "regular" cores + HT). For basic day-to-day work, they both felt the same.
As you say, I'd expect a part with more P cores to be interesting for my use case, which is exactly the same as your sibling's [0]
i don't play videogames and i don't do ai/ml, video editing or whatever.
i spend most of my time in firefox and in terminals. i do use virtual machines quite a lot to test stuff around, so high cores count and high memory would be a plus for me.
but an nvidia gpu is a deal breaker for me. I just don't want it. I don't need it, i don't use it, it just makes the whole thing less usable. and it draws a lot power that i'd prefer using otherwise.
the intel integrated gpu is not only sufficient for me, i actually WANT it. it just works under gnu/linux and i don't have to mess with drivers.
I cannot stress enough how much i hate having to deal with proprietary drivers.
and last time i played a videogame, it was openarena, and it ran beautifully on the intel hd 600 my work laptop had (dell latitude 7390, great little machine).
edit: regarding amd gpus... i'm not sure. i've been told they work with open source drivers, but still it's a power usage i would happily avoid.
I mean, in a laptop, you dont use the dGPU on battery unless its for compute, period. It should just be sitting there turned off (which indeed means you dont want it). But if its sucking any power, thats a bug.
That being said, Firefox and Chromium use the integrated GPU more than you think, and they feel faster with a stronger IGP. Just try disabling some of the gpu acceleration and see how it feels.
That is not how GPU multiplexing works on Laptops. My dGPU most certainly works on battery, though I can directly disable it if I like (or use a power saving power plan).
It works, but it sucks power like no tomorrow if it actually stays on. Its basically unusable if you want more than hour or two of battery life (or you use it in short bursts for compute).
For example, I could disable the nvidia card on my ThinkPad W530, but then i'd lose the use of the external video output (because it was wired to the nvidia card).
and yes i know it depends on whether the laptop has a mux or not... it's not always easy to determine, not all reviews go explicitly over this detail...
I just don't want a discrete graphics card. Not an nVidia one for sure.
> That being said, Firefox and Chromium use the integrated GPU more than you think, and they feel faster with a stronger IGP. Just try disabling some of the gpu acceleration and see how it feels.
i have strongs doubts the doxygen webpages i spend time on will get any faster with an nvidia 3090.
I haven't noticed any difference between an i5-8500 with its integrated GPU and my gaming PC with an RX5000 something. Both running latest windows and edge.
They actually feel similar to Firefox running on Linux on an i5-6500's IGP.
It probably depends on the websites you visit. The only thing better on the newer ones is video decoding for YouTube or the like. I usually actively avoid websites with animations and other similar stuff as I find them very unpleasant to use. But I've noticed those typically tend to use the CPU, even on a Zen3 or Intel 11th gen IGP, even on Windows + Edge.
On my nvidia thinkpad I have had to completely disable the integrated GPU in bios and use discrete only due to constant bugs with hybrid graphics on Linux. This means my battery life is less than two hours.
You can usually disable it completely in the BIOS, and it won't be drawing power if it isn't doing anything. I can get wanting to not pay extra, but realistically supporting an extra custom hardware configuration probably costs more than they would save on the part.
> The issue is that on some machines, the video-out ports go through the dGPU. Does that still work if the card is disabled in the BIOS?
Traditionally yes; the machine is designed to only run the dGPU under heavy load, not all the time, the BIOS setting just disables doing the thing that you did in the "heavy load" condition. I guess there might be machines it wouldn't work on but I've never known any.
The Radeon 680M/780M compete pretty well with base M1/M2 GPUs. And there are going to be some 7040 series chips with them like the 7940HS. No, it's not a 16 core monster but I imagine Aya Neo will shove it in a handheld in like 5 months.
LG Gram is a multipurpose laptop. You can use it's screen also as a mirror and keyboard has so interesting design you can play twister game with your fingers whether you like it or not.
It would be nice if we could get non-multiplexed edit keys in the unused space. That is the biggest failing of modern laptops. Massive horizontal real estate and they're packing keys in like it's 1995.
Previous framework was 3:2, not 16:9 so there isn't a ridiculous amount of width. But yes, a vertical column of dedicated keys is a decent compromise that doesn't shift the keyboard over much. I had that on an older luggable.
You don't have to shift the keyboard over at all. That can stay centered and there's room for a 2x3 block of edit keys plus an inverted-T with full height up and down arrows. All it takes is willingness to be a leader instead of a follower copying one company's misguided minimalism-at-all-costs.
Because of the screen ratio, there's simply no space to do that on a Framework 13, without reducing key spacing: there is almost exactly 1 cm of space between the last keys and the actual edge of the frame.
On the Framework 16, I wouldn't be surprised if someone makes that: it looks like the design would accommodate a left/right-hand side narrow input module with the keyboard still centred.
I'd prefer inverted-T with half height keys and use fn as a modifier for pgup/pgdn/home/end. Don't really need insert or delete though, but there's already a key for that.
Maybe you mean an 85-key "75% keyboard". It's a popular request on the frame.work community discussion of "The keyboard".
On keyboard layouts like Framework's, I still have trouble trying to touch-type the right control key with my ring finger-nail. I prefer a "75% keyboard" with an navigation/editing column which provides room for the right ctrl key further right. (And full size arrow keys. And makes shift-End easy for selecting rest of line.)
Maybe we can hope a Kickstarter keyboard maker will make a Framework version in a few years. Main limitation is the thinness, only 3.7mm.
If they're using qmkesque firmware you could start programming some interesting new layers to give you functionality like this. I used to require a full sized keyboard so I'd have all this functionality like home and end keys and stuff, and then I started using keyboards with qmk and similar, and now I have all sorts of layers and macros and etc and I seriously can never go back.
I tend to favor left control, and right shift myself... discovered this using a KB where the up arrow was at the edge of the right shift, most annoying keyboard to type on ever... all of the sudden my typing is in the middle of already typed text. If typing while looking at something else, I'd see a garbled mess too often.
Do you have very large hands or a narrow keyboard? On a normal 85-key "75% keyboard", I cannot reach the up-arrow key with my pinky while the rest of my fingers are on the home keys.
I've had similar cursor problems with those keyboards that have page-up and page-dn next to the up-arrow key. Just a little finger displacement when trying to left-arrow and the cursor is suddenly inserting my subsequent typed text far from the intended location.
Also height of keys. I don't understand the rush towards thinner and thinner that killed the more reasonably higher-keys keyboards.I don't need a 1mm laptop, I need a laptop I can comfortably type in.
The customization goes further than just yes/no to the number pad. Looks like a whole "Input Module system" that could open up opportunities for different keyboard designs (DVORAK, various locales, maybe different switches?). I love this idea, excited to see what folks do with it. Makes me think of Apple's Touchbar nonsense -- why force a number pad, or a second touchscreen on folks who don't want one? Such a good iteration on the Framework idea.
They're using QMK on their prototype keyboard and numpad input modules (which are RP2040-based internally).
The interface for input modules to the rest of the computer is "just" USB 2.0, so you could put just about anything there that'll fit in one of the input module form factors. Although high resolution, high frame rate touchscreens are probably out because of bandwidth constraints (since it is only USB 2.0).
It's plausible to use something like a DisplayLink controller, or for lower resolution even a microcontroller with flexible high speed peripherals like an RP2040 with the PIO interface.
Given that there's just more room in the 16", I wonder how hard it would be to just build an adapter board (and some 3d printed scaffolding) to just mount an actual lenovo thinkpad replacement part...
When it comes to the keyboard, the room that matters is the height, and both the FW 13 or 16 have a maximum height of 3.8mm
Unless one can come up with a way to produce a custom keyboard with the same height restriction, using something like a Thinkpad keyboard which is considerably taller would essentially require a custom taller chassis.
Split ortholinear keyboard with a clit mouse. If someone makes this to where I can just plug and play and program the keyboard firmware directly I'm switching at the next opportune moment(I'm not a hardware guy, I stop at the lowest levels of programming and leave the soldering to steadier hands...)
Though my Thinkpad can still handle another 5 years of wanton abuse I'm sure, so I have all the time in the world.
Man my next laptop is gonna be awesome. Literally everything I want in a laptop, which has never existed until now(well, soon).
I'd love someone to do a split ortholinear keyboard on a laptop. That'd probably push it to instant buy for me almost regardless of the rest of the specs etc.
Split ortholinear is possible. We didn't show this, but it is technically possible to build an Input Module for the Framework Laptop 16 that is full width, giving a lot of horizontal real estate for a split layout.
Please do an ortho or split ortho layout. I get that it'd be low volume but I'd be willing to pay a premium if it meant being able to keep my Helidox Corne at home while I travel or go into the office, without sacrificing ergonomics.
Looking forward to purchasing the 16 regardless though. Keep up the amazing work / mission with Framework. Fantastic work to the team.
As a dev with hands that won't let me use a standard laptop keyboard for more than a few minutes without being in pain, this is so needed. It would be so good to be mobile again.
Have you had much experience with mechanical keyboards? I do much better with just unicomp m-style buckling spring or cherry mx brown (real cherry, not the fake that some use).
There is a huge untapped market segment here. It's only a matter of time before someone makes it happen.
Honestly, I see the biggest problem that people have entering the ortholinear/ergonomic keyboard market is cost: the obvious choice is to target mechanical keyswitches, but you quickly realize your consumer base would rather pay more (in parts cost and construction effort) to get exactlty what they want.
But what if you went as cheap as possible instead? What if you made flat rubber dome boards that cost next to nothing to manufacture; and sold them for $40 a pop? You would probably broaden your target audience by a factor of 100, and have huge profit margins.
Doing that for an integrated laptop keyboard would be even easier to sell. I would gladly use shitty rubber dome switches if it meant having a split ortho layout that I didn't have to lug around separately from my laptop. I would buy that yesterday.
Framework should beg Sensel for a blank touch surface...which you could press a couple of macro keys to entirely transform the surface with.
users print out a custom layout to guide their fingers.../dream
Agreed, I'd kill for a modern version of the TouchStream LP. Comfiest and quietest keyboard I've ever used, with no elbow strain from reaching for a mouse/trackpad. They even had a product that could replace a MacBook keyboard. [1]
That being said, the company was acquired by Apple and turned into the iPhone and iPod Touch. The keyboards were effectively museum pieces at that point. I wonder if the patent situation would allow for third parties to revisit the design. It'd be interesting to see what could be done with by borrowing ideas from the programmable keyboard community and mobile touch hardware.
A standardised laptop GPU connector could be a game changer and would be a killer feature. There have been a few false starts in that space with MXM, but they never really took off.
It’s often the first thing people upgrade on a desktop (it’s much more impactful than a RAM and CPU upgrade over the same timeframe), but is rarely an option for laptop users.
Wow that's the fastest I've ever seen a site go down. Site was vaporized as soon as the announcement was made.
I'm still happy with my 11th gen framework. Anybody who's looking into upgrading please consider re-using your mainboard. They have 3D printing schematics you can use to turn it into a mini desktop PC!: https://github.com/FrameworkComputer/Mainboard
tl;dr: A 16" laptop with modular dedicated graphics (PCIe x8 either v4 or v5 with fans on the module), expansion bay cards (dual m.s SSDs), a new 'input system' with multiple keyboards in multiple languages and optional RGB lighting as well as support for numpads and secondary displays. Target launch is 'late 2023'.
While I'm glad they are actually making a product for their actual niche...
Personally, looking at the 16...it's just a 'You were supposed to be the chosen one.' gif.
Damn near everything is proprietary, even if they were to open source it, they have basically created their own new standards for everything from the GPU to the new RGB modules that seem like insane distractions. Also those trackpad spacers are a solution but yikes that looks like trash.
All I see is SKU overload and inventory problems all while they are distracted with solving RGB for some wild reason.
MXM is an 'open-ish' standard that others build for. My real weird view is that they have really went a different direction and their GPU upgrade has a whole rear deck you install and all that rather than building it into the product. That deck will only ever work with the Framework 16.
Like, this isn't a standard at all, it's just open for Framework to gain from. It reminds me of Google's Project Ara but for the wrong reasons.
MXM is alright for upgrading, but it's not modular the way this aims to be.
This includes cooling, so a small GPU could have smaller heatsinks and a beefy GPU could have huge heatsinks.
And if you don't want the GPU and the associated power draw plus the size and weight of the heatsinks, you can swap it for another module if you don't want to game on the go.
As they said in their presentation, it's not just for GPUs. You could have a supplemental battery. They will offer a storage expansion that gives you more M.2 slots. Hypothetically you could even just make a module that lets you fit a desktop GPU on the back, if you provided external power.
I love how this launch included new reuse paths for old components, including a new mainboard case [1] and battery / screen case designs. If Framework maintains mechanical compatibility between generations, things are going to get really interesting.
Didn't use a guide, it's a bog-standard Arcade1Up conversion. A couple years ago I looked up a display driver board providing an HDMI input to the stock display and bought a $50 Amazon-sourced EG STARTS-brand cheapo kit of sticks and buttons with USB encoders. They were drop-in replacements for the stock deck's buttons and sticks. The emulator driver right now is just a 6th-gen i5 HP EliteDesk Mini fleet-salvaged workstation booting into RetroArch.
These days you can go even easier with fully installed plug-and-play Arcade1Up control panel replacements that are literally drop-in - MDF, sticks, buttons, encoders, and audio/display driver card are all built in. You just pull the stock control deck out, drop the $150 Intec deck in, and connect whatever PC/console/FPGA into the video in and USB controller outputs that it provides. You can build a cab around one out of plywood or just grab a random Arcade1Up on sale somewhere.
I'm also intrigued by the business opportunities of providing more ways for fans to spend money than just buying a new laptop every 4 years. If I can upgrade to a new mainboard and turn my old one into a functional desktop, I'm much more likely to add in a $500 mainboard upgrade at the 2 year mark.
I'm not the parent, but they've announced they've partnered with Cooler Master to release a case [0]. A while back they released drawings so you can 3D print one [1].
I'm using this 5-bay USB 3.2 drive enclosure [1] and the official Framework 3d printed case [2] in polycarbonate ordered via Craftcloud (the Cooler Master case would have been both cheaper and easier!)
I also looked at TB4 disk enclosures, but the bandwidth was overkill for spinning rust hard disks. There's a lot you can do with 4 USB-C/TB4 ports. I also have a Coral USB stick connected for frigate.video inference and a 2.5Gbps USB NIC (waiting for better thunderbolt 10Gbps NICs to become available). One fun convergence is when I want a keyboard / display / mouse I can plug the mainboard server into the same thunderbolt dock I use for my laptop peripherals.
Hopefully a matte screen option??? That's the one thing holding me back from getting one. Those stickers they offer don't cut it. Amd vs Intel doesn't really bother me either way but a matte screen is an absolute must.
Don't have time to check now with the site being so slow but will do later. Good to see the amount of attention this receives though.
I very much much agree with this. Laptop keyboards are incredibly painful, and I can't use a laptop because of this.
a split keyboard would solve my issue and I'd definitely buy the laptop.
There is something charming about the amateurish feel of their livestream, like the money I've given them is going to real people doing real work and not into a bloated marketing budget.
Not just charming, but also makes me feel like they value my money and want to make sure it goes toward the mission. It's easy to say "hire a production company", but personally, I can put up with a clunky PR event if they use the savings to keep pushing the mission forward.
> Coming later in 2023: Belgium, Taiwan, Italy, Spain
Does anybody know why they do this country by country roll out? I'd be perfectly happy to order a US layout laptop but they don't allow me to place an order.
It's probably harder being a US company selling into each EU country than it is for an EU-based company to sell to the rest of the EU and also the US, too?
The alternative would be not selling directly to consumer, using some intermediary like Amazon or whatever.
I don't think is that harder. Also, my Framework(bought in Austria and than shipped it to Romania - still in EU) was actually sold by a company registered in Germany.
Maybe they'd have a lot of extra returns if people from these countries didn't pay attention and bought a laptop expecting the usual keyboard layout for their languages.
I don’t think so (not mentioned during the talk, so probably not), but the input system is being opened up (with both hardware specs and some their keyboard software released), so it should be possible for a third party to do that.
I would probably switch after decades of thinkpads (and one Mac I just didn't like) if they offered a model with AMD chip, a good keyboard, Trackpoint, and OLED display.
Hmm there are quite a few 16" OLEDs on the market these days, I assume there is a common size for '16".' I'm hoping not to upgrade for another year, so will see then.
IBM's trackpoint patent expired in 2017. I have no idea why there hasn't been a custom keyboard with a trackpoint out yet. Seems like that would be the first thing hacker-types would add.
The way the numpad is fully optional on the 16" is great. I get that numpads are non-optional for some but for my usage it's mostly a nuisance, which makes the non-optional numpads on a lot of laptops a near-dealbreaker.
For me they were completely useless too but I've remapped mine to a 3x3 desktop switcher and some other hotkeys on KDE. Now I can't do without it, it's like having 9 computers in one <3
I wish Framework had more color options in their Marketplace (e.g gold, torquise, pink), especially considering how they advertise themselves as being modular / customizable.
We're adding more Bezel colors throughout the year. We have Transparent available now for pre-ordering. We've improved our manufacturing infrastructure for Bezels recently to make it much easier to create new colors than it was previously.
Hey Nirav, thanks so much for the progress you and your team have done! Just had a question if there's been some benchmarking done for Linux on the new AMD and Intel mainboards.
We don't have any benchmarks we can share yet, but we absolutely expect that there will be press reviews that do comparisons between Intel and AMD-based Framework Laptops.
As one of the QMK Firmware maintainers, it's great to see you're going ahead with the customisable side of things even on the input side.
Might I suggest engaging with QMK early so that we can avoid the usual manufacturer "hey can you merge this, we've shipped already and forgot to raise the PR until now"?
Absolutely! We’ve been chatting with Jack Humbert, and he’s actually been prototyping an ortholinear keyboard module for the Framework Laptop 16. There are likely some photos of it in press articles about the event today.
That's great news, any chance for an orthostaggered module as our fingers have different lengths for a linear one (or maybe that's what you meant with just terminology slip)
If you folks and Jack manage to make the OLKB module a reality, you'll have me vendorlocked for life. I think I'd go so far as buying 3 of those modules just to be extra sure that I'm set with replacement parts.
Uh... completely unrelated question: have you laid out any plans regarding module backcompatibility on future boards?
They do use different mainboards in order to enable all of the functionality that exists in the Framework Laptop 16 (6x Expansion Cards, the Expansion Bay, and the Input Module system).
I didn't quite get what form the modular upgradeable graphics would take:
* Is this an "egpu connected by a cable" situation, or something actually integrated into the laptop's body?
* What are realistic possibilities for how much gpu, and how much of a premium would it be over other laptop or desktop gpu's? Is this something you would be making, or third parties?
> The Framework Laptop 16 has an Expansion Bay system that contains both the cooling fans and an interface for high-power, high-speed PCIe peripherals like GPUs. Expansion Bay modules can extend in both thickness and depth to accomodate higher power requirements.
Hey! I pre-ordered an AMD version as soon as the site would handle it, but I wasn't able to choose between the mattte and glossy display. Is there any way to get the matte display with the new AMD option?
The Matte display version is pre-installed on all 13th Gen Intel Core and AMD Ryzen 7040 Series Framework Laptops. If you do want a glossy display, it's available as a module in the Marketplace to swap in (or you can pick up an older 11th Gen, 12th Gen, or Chromebook Edition which also have the original display).
It's ultimately inventory costs and supply chain complexity to manage a larger number of laptop SKUs (CPU options * DIY/pre-built * keyboard options * any other configurable variable). We're a small team, so when there are areas of the product we can condense into having one default configuration, we take those efficiencies, while still enabling an "escape valve" through the Marketplace.
I understand the supply concerns, but in the future please consider a glossy display as default option (even if just in the DIY edition). It seems wasteful to have to buy matte display only to replace with a glossy one. A lot of people in the monitors community hate matte displays and actively avoid buying devices with them.
Looking at your mechanical design for the large input module, though I can't compare the height at the moment, it appears to be almost the same size as the keyboard on the Framework 13. While it would obviously require a new input cover design, would it be possible to design an assembly that would allow a single large input module to be used as the keyboard on a Framework 13? Or would there be height problems?
The key structure actually is the same as the Framework Laptop 13, including the same key travel. The overall module is different though to handle the interface differences, in a way that makes it not possible to retrofit into a Framework Laptop 13 unfortunately (we thought about this, but couldn't make it work well).
Is there any progress being made with some of the nagging issues that impact the 12th gen intel laptops when using Linux?
Specifically poor battery life (especially during standby when using the USB-A module) and not being able to use the brightness keys unless you disable the sensor?
Our latest beta BIOS updates improve the power consumption while using USB-A Expansion Cards on 11th and 12th Gen. We have some additional improvements we’re working on around this on HDMI and DP cards as well. For the brightness key/ALS conflict, this is pending additional driver work.
Are you talking about the one from December that's in beta[1]? Do you know when it'll be ready for primetime? Sounds like it turns off thunderbolt on the left side of the laptop so I've been reluctant to try it
Any chance you will introduce a "good keyboard" option? Like old-school ThinkPad keyboards. (key travel depth, figured keycaps, maybe additional row of keys)
Will it be possible to upgrade my current framework laptop to the new 16" form factor? It feels like it should be a possible upgrade, but I haven't seen anything saying that it will be possible, so I'd like to confirm.
i really hope the battery is much better than the pathetic 3 hours i get from my current one. as im writing this, ive used my laptop for around 3 hours, 20 minutes was to run a kvm and the rest is just a browser, and im at 18% remaining.
i might have said before that i regret getting the framework, i might think about upgrading my mainboard and battery if it turns out to be working well. key word "maybe"...
490 comments
[ 2.2 ms ] story [ 319 ms ] threadEdit: Also the input deck and the ability to attach an egpu as well as upgrade to a newer egpu in the future are huge. This is a fantastic leap forward for Framework.
Jim Keller should make a Tenstorrent board. They already want to make desktop cards, and that would be a great way to market them.
https://morethanmoore.substack.com/p/interview-with-jim-kell...
If he quit Tenstorrent, that must have been very recent.
Not only would they get exposure, but they would love to see Framework-owning devs experimenting on Grayskull like its an RTX 3060.
Of course it would probably be linux only, but still...
Anyway, excited for a Framework version too! While I prefer AMD integrated graphics to Intel, NVidia dGPU would be even better, and even better still if it was upgradable. Are laptop GPUs still available on MXM cards?
Btw, this is excellent advice. Funny story, I have a FreeNAS device and the motherboard died and I thought "Oh my, I need to bring my ZFS volumes up on another machine, but I didn't have another machine with 6 SATA bays! I ended up having the drives all sitting out on the workbench connected a mainboard with 8 SATA ports so that I could create an archive of the data, and then got the mainboard fixed so I could re-assemble and re-use the FreeNAS but still it alerted me to the fact that I really needed a 6-8 drive cabinet if I wanted to do this again.
Hardware raid is legacy :)
Sure, for personal usage there is almost no usage for the HW RAID, but when you need to make sure what the system would always boot and it can't be serviced in hours/days - then you have almost no options for SW RAID.
No problem to put /boot on a raid1 on a small partition across all drives, so that any drive can boot, and no problem to even include a whole self-contained remotely accessible recovery os. It's a little bit more work to set up, but if you are professing a need for that, then a little extra setup is de rigueur. I remotely administered a ton of linux boxes in racks scattered across the US like that for years. Although I had out of band serial console access and could do full bare metal reinstall that way, I could also do it from any neighboring machine that was still running in the same rack if I had to, with a combination of network booting and/or booting from any one of the normal drives normal raid1 copies of the /boot partition.
Further remote-able fallback options that I never even had to use but could: Local hands just plucks a hot-swap drive from any of my other machines and pops it into the bad machine. All drives had the same bootable partition and all drives are redundant and so they could yank literally any one from the wall of server fronts. Or, better, local hands just plugs in a thumb drive and I take care of the rest. Thumb drive is already sitting there for that purpose, or they could make a new one from a download. But with 8 to 24 hot-swap drives per machine, meaning 8-24 copies of /boot, I never even once needed local hands to so much as plug in a thumb drive.
There is just no problem at all with sw raid. It only provides options, not remove them.
Did you even read my comment? It's quite clear what your environment was in the data-centers, with spares and remote hands.
Mine wasn't and then I say three months I don't kid or jest.
> No problem to put /boot on a raid1 on a small partition across all drives
This is exactly the problem. If the drive isn't totally dead (like it doesn't even respond to IDENT) then there is a chance what the BIOS/UEFI would try to boot from it and even succeed in that (ie would load the MBR/boot app) and then there is no way to alter the boot process at this point. HW RAID card provides a single boot point and handle the failing drives by itself, never exposing those shenanigans to the upper level.
Like sure, you are happy with your setup, you never had a bad experience with it, you always had OOB management and remote hands - but it doesn't means what it a silver bullet working 100% of times for everyone.
Yes, I've seen systems with SW/fake RAID failed to boot because the boot process failed after selecting a half-dead drive as a boot device, with my own eyes. Thankfully I was geographically close to them, not 5000km away.
Yes, I serviced and prepared systems for the 5000km away divisions and they are really serviced only a couple of months in a year, all other time you need an extremely urgent reasons why do you need to a rent a heavy 'copter to go there. No, there is no remote-hands there. The maximum point of IT-competency there is raking bills with satellite Internet.
It’s probably time for me to figure out converting over to software raid.
Thanks for the nudge.
I bought four cheap SATA-USB3 adapters, plugged them into two USB3 hubs, which I then plugged into the two USB3 ports of a Raspberry Pi 4, and arranged it all quite precariously in a small cardboard box that I cut holes into for airflow. Performance was terrible, of course, but it worked well enough until I could build a proper new NAS box.
With generic software raid (mdadm), even if you only had a single usb port and a single internal drive, you could image all the drives one at a time and then access the array of images on the single big drive. (not uncommon since usually time has passed by the time something fails)
It also goes the other way if you needed it to. Say a single array member was 4T but you only have a bunch of 1T drives, no problem, you can assemble 4 1T drives into a 4T container to hold the 4T image, and then use that image as an aray member itself.
Even if you don't have any loose external drives, you could even do it all via network shares with pieces residing on all of your other family members laptops and desktops, or every desk in an office, while they all still continue running windows and doing their normal jobs I mean.
Some of the possibilities are slow or fragile or both, so of course you don't set out to use 12 usb2.0 ports, but the point is essentially anything is possible, and you don't have to worry about predicting or planning for every possibility, you just don't have to care about how you'll recover the array in the future because it doesn't matter what form storage takes at that time, or what form you happen to have available. It would almost never make sense to do some things, but the point is that mdadm just doesn't care.
For a machine with only 2 or 4 internal drives where you want to use raid0 for max throughput, and don't want to rely on any special firmware support for booting raid0, just partition the drives so that /boot is a small raid1 across all the same drives, so that any of the drives could boot. Bonus, it automatically makes all the members of your main raid0 slightly smaller than the drive's nominal size, which means you can always fit them onto some other replacement drive later, even if the different manufacturers count bytes and formatting overhead differently.
I come from the days of sco unix on scsi hardware raid with full featured expensive cards and I do not miss it.
[0] https://www.cgsecurity.org/wiki/TestDisk
On the Asus, are both screens touch screens? Or is only the lower screen a touch screen? I had looked at 20:2 type touch screens to do a sort of "media bar" setup on my desktop but didn't find anything at the time I could use. I'm wondering how well such a setup might work.
If only I'd waited an extra month to order my new work equipment.
People will still want different screen resolutions, aspect ratios, keyboard layouts, etc. Now that AMD is available people will want particular CPU SKUs. It will never end.
I wish Framework luck, because they are targeting the pickiest userbase that exists.
But yeah, when I was suddenly in the market for a new laptop anyway, the extensive freedom to upgrade and configure this made it an extremely easy sell compared to anything else that would have been on my radar.
I suspect the choruses will be much quieter now. If I hadn't already landed on a Beelink GTR6 mini-PC desktop + MNT Reform laptop, I'd be considering this finally - indeed, one of many reasons I went Reform over Framework originally was because Framework was married to Intel. And frankly, 3:2 still makes Framework tempting despite my lack of need for it.
Definitely going Framework for my next laptop in a couple years... really tempted to get one sooner than later and pass down my M1 air to my daughter.
Too many platforms lock users into decisions they think are best for them, or worse, that they think are best for the company making said platform. It’s refreshing to be given not only options, but options I actually want.
But some people just like supporting the underdog.
There's a few instances where AMD GPU drivers even outshine their windows counterparts like in OpenGL performance. Intel's driver support is great as well, but their performance recently has not matched AMD, and their iGPUs don't compare to AMD's iGPUs.
Also for beefier laptops with discreet GPUs, intel up until very recently would need an NVIDIA card to get decent performance which would require NVIDIA drivers, which are notorious for complicating things on Linux. On the other hand AMD laptops with discreet AMD gpus already have their drivers built into the kernel which removes almost any potential complication.
Also there's personal preference, which in my case leans toward AMD because of their generally more open stance with their new technology, like raytracing and FSR which is contrasted by NVIDIAs typically closed off approach to the same things. Intel isn't as bad as NVIDIA in this department either but this is just personal opinion.
Lastly AMD CPUs are generally cheaper than intel, and in the last few generations performed better overall than intel.
TL:DR AMD on linux has great drivers with good support, the synergy between AMD's CPUs and GPUs and their combined APUs are very good, and personally I appreciate AMD's more open approach to technology like FSR, even if they aren't anywhere near perfect.
AMD still tends to have lower actual TDP than Intel, which is more efficient in a laptop and easier to cool in a desktop.
If you're going to use the integrated GPU, RDNA generally performs better in games.
My home linux machine is a ninth gen intel with Nvidia. Its technically a notebook, but more like a portable workstation. Its powerful, but battery life is terrible (<3) and you can hear the fans. It can game quite well however.
The reason that people want AMD CPUs is simple - they’re much better than what Intel is offering. We can see empirically how wide this gap is now since we have independent reviews of identical laptop platforms from Lenovo and HP’s ultrathin business laptops (the same segment as Framework):
HP EliteBook 840 G9 1280P vs HP EliteBook 845 G9 6950HS [1] - these are both respective flagship parts, and we can see the AMD version has both a +23% performance rating and a +23% better battery runtime. Note, that on the HP website currently, the same configuration AMD version is also >30% cheaper.
Lenovo ThinkPad T14s G3 6850U vs 1235U [2] - since the Intel version reviewed isn’t a top of the line model, it’s not fair to compare performance numbers, but even with the significantly lower TDP Intel part (15W vs 28W), the AMD version of the laptop ends up with an even bigger lead with a 35% better battery runtime. The AMD version is also 10% cheaper than their Intel counterparts from the Lenovo site pricing.
In both these cases, the AMD version wins significantly on processing power, battery efficiency and price. The 7040 should extend the lead on the latter, and that's not taking GPU performance into account [3], where the AMD Radeon 680M simply crushes the Intel Xe 96EU by an average of +88% in game performance, and +135% in synthetics (and again, the 7040’s RDNA3 GPU and Xilinx AI core will extend the lead even more for the upcoming generation).
[1] https://www.notebookcheck.net/HP-EliteBook-840-G9-laptop-rev...
[2] https://www.notebookcheck.net/Lenovo-ThinkPad-T14s-G3-AMD-la...
[3] https://www.notebookcheck.com/Iris-Xe-G7-96EUs-vs-Radeon-680...
USB features depend ultimately on the laptop manufacturers' implementation (eg neither the HP or Lenovo models linked implement USB4 but they are in other models like the Lenovo Z13), but I haven't seen any that claim to have DP-alt support not be able to support it. As the Framework will have USB4 (and DP-alt and PCIe encapsulation are mandated by Microsoft for Windows 11 compatibility), I don't think that's a real worry for anyone.
(worth the down votes)
I really don’t need another notebook, but I like what Framework is doing, and I’ll be watching the reviews.
The only question I'm left with now, which Framework to get? Do I pre-order the AMD 7040, or do I wait for the new 16 so that I can get a GPU? I'm thinking about going the AMD route now, buying the dGPU option they have and using it as an eGPU when it becomes available, and then switching to the 16-inch chassis when I'm ready...
The 16" will provably perform better though, if you can wait.
https://www.notebookcheck.net/AMD-s-Radeon-780M-iGPU-beats-t...
I also think it's important to not underestimate the flexibility and marketing aspects of the eGPU design they went with. Someone who has a laptop they use for playing games with a still fine CPU but outdated GPU might look at this and just buy the eGPU. Then when it comes time to upgrade the laptop itself they might be all-in on the value of an upgradable system.
The beauty of this is they don't have to choose. The GPU can be anything that communicates over Thunderbolt and fits in the physical module specs.
In other words, only the GPU modules that Framework chooses to make.
If the framework laptop grows in market size that there’s enough demand for it, then anyone could manufacture and sell those components, which is kind of the point.
I think it is pretty clear that Framework intends to allow you to buy the most barebones laptop you can. I mean they let you buy 0 of the little cartridge modules, I assume they won't force people to buy one of these PCIex8 modules.
With this in mind, trampling module vendors doesn't seem like a great idea. What's the point of creating an ecosystem if there's nobody around to participate?
Festival and concert tickets are frequently done this way and the queue renege rate lets you model how many you would have lost if your site failed to load and at what duration.
https://us.starlabs.systems/pages/starfighter
TBH I am kind of shocked Framework picked up the AMD 7040 series so quickly. Usually small manufacturers are stuck with older gens.
Starfighter doesn’t have the proven “longevity” and true commitment to mission that Framework has had time to demonstrate, now through its second major release.
That all said, Starfighter looks like it has some really cool features between open source boot firmware, removeable camera and Wi-Fi kill switch.
If you want a better or more libre Linux laptop right now, get a Starfighter. If you want a laptop with more hardware hackability and replacement options, at the expense of libre/privacy features, battery life, supported Linux options, and high-end performance, Framework's an option.
And is hardware the limiting thing here?
Not the terrible arm OSes?
https://arstechnica.com/gadgets/2022/02/lenovo-announces-the...
ARM is great in theory, but I don't see any SoC on the horizon that I would really crave. This includes Apple, as I am either stuck with OSX or (at best) have most of the interesting bits of the SoC nonfunctional in linux or Windows.
And they support other stuff!
I wish the framework site was up, as I want to see what this specification looks like. If you look inside an Asus Zephyrus or whatever, the cooling setup needed to support such a thing in a thin laptop is mad. The 13", for instance, has a pancake sized 2D headspreader + liquid metal TIM, which very much goes against repairability. The 16" is a wall of heatpipes. I wonder how framework is addressing this.
EDIT: specification is here, but down at this moment: https://github.com/FrameworkComputer/ExpansionBay
It is bulky though. Moreover, this bulk is not enough to keep the keyboard from getting really hot under load.
- New Framework laptops are available using Intel 13th Gen and AMD 7040 series.
- New 61Wh battery, new matte display option & a Cooler Master mainboard case for $39.
- New 16 inch Framework laptop. Allows you to have a number pad or not, up to you and customizable. If I understood the livestream correctly, there are upgradable dedicated GPUs that connect to the 16 inch laptop.
This sounds fantastic, I very well may buy a Framework laptop later this week/weekend.
EDIT: Their site came back up, after reading the blog post I pre-ordered the Ryzen 7 13" to be my new Linux laptop.
"For the AMD Ryzen™ 7040 Series processors, we’ll be sharing more detailed specifications as we get closer to shipment."
Source: https://frame.work/gb/en/blog/framework-laptop-13-with-13th-...
On previous generations: the 4000 series starts to hit a relative performance wall at around 35W. The 5000 series clocked higher and hence gained more moving from 15 to 35 and 35 to 45.
Not sure about the 6000 series, but you can see for yourself: https://www.notebookcheck.net/AMD-Ryzen-7-6800U-Processor-Be...
Another point: you can throttle the 35W HS series to 15W, and get the laptop to run fanless or almost fanless since it is designed to cool a higher TDP.
Shipping estimate is Q3.
I'm guessing they don't want to officially support any specific distro, and offering support for multiple distros is probably a bit outside their wheelhouse at the moment. [e: this is perhaps not the case]
https://frame.work/linux
> Officially supported
> Fedora 37
> Ubuntu 22.04 LTS
https://frame.work/blog/framework-laptop-13-with-13th-gen-in...
Optimized for Linux
We continue to focus on solid Linux support, and we’re happy to share that Fedora 38 and Ubuntu 22.04 will work fantastically out of the box for both the 13th Gen Intel® Core™ and AMD Ryzen™ 7040 Series models. We have thorough setup and troubleshooting guides and will continue to provide official support for these two distributions. Manjaro XFCE 22.0 and Linux Mint 21.1 are also working great, and we’re detailing the documentation for those as well. You can check compatibility with popular distros as we continue to test them on our Linux page or in the Framework Community.
I might find myself ordering one of these.
Not shipping with Linux installed gets Framework out of the business of managing distros and images themselves. They also dodge a neverending barrage of "But what about my favorite distro X?" that would surely come for them as soon as they picked some fixed set of distros.
If you are uncomfortable assembling computer components and installing an OS yourself then linux is probably not for you. This acts as a filter for people who are serious.
It's a lot easier to support newbie computer users on windows, and they are a lot more likely to get support from friends and family.
They would be shipping a DIY edition anyways, this vastly simplifies their inventory.
These may offend your Linux-maximalist sensibilities, c'est la vie.
Personally, I kind of wish System76 would sell framework laptops with a generalized configuration and PopOS pre-insalled.
Including any linux distro would be better, because it would show a growing linux desktop market share.
See also https://gist.github.com/Informatic/49bd034d43e054bd1d8d4fec3...
In the second case, you have to pass through SLIC and MSDM ACPI tables from host to guest. For example, like this: https://gist.github.com/Informatic/49bd034d43e054bd1d8d4fec3...
The 7040 series is good, but its not dGPU class like an M1/M2 Max. I think AMD has been scared to make one because OEMs won't want it.
But I do use the CPU from time to time, so a beefier part would be useful. I would rather not pay for a dGPU which I'd use maybe 1% of the time, and which would cost more and weigh more and possibly be a PITA to manage under Linux.
But what it sounds like you want is an E-core mad part from Intel. I always thought they should sell a 20-30 core laptop part with only one or two big cores as a compilation monster.
Indeed. I've been looking at the 12th gen laptops we're getting at work, but those were underwhelming. I don't remember which model CPU it was, but it only had 2 P cores (those are basically for "office work", so no one cares about performance). For compilation (Rust) it was in the same ballpark as my 11th gen i7 (so 4 "regular" cores + HT). For basic day-to-day work, they both felt the same.
As you say, I'd expect a part with more P cores to be interesting for my use case, which is exactly the same as your sibling's [0]
[0] https://news.ycombinator.com/threads?id=vladvasiliu#35280246
i spend most of my time in firefox and in terminals. i do use virtual machines quite a lot to test stuff around, so high cores count and high memory would be a plus for me.
but an nvidia gpu is a deal breaker for me. I just don't want it. I don't need it, i don't use it, it just makes the whole thing less usable. and it draws a lot power that i'd prefer using otherwise.
the intel integrated gpu is not only sufficient for me, i actually WANT it. it just works under gnu/linux and i don't have to mess with drivers.
I cannot stress enough how much i hate having to deal with proprietary drivers.
and last time i played a videogame, it was openarena, and it ran beautifully on the intel hd 600 my work laptop had (dell latitude 7390, great little machine).
edit: regarding amd gpus... i'm not sure. i've been told they work with open source drivers, but still it's a power usage i would happily avoid.
That being said, Firefox and Chromium use the integrated GPU more than you think, and they feel faster with a stronger IGP. Just try disabling some of the gpu acceleration and see how it feels.
For example, I could disable the nvidia card on my ThinkPad W530, but then i'd lose the use of the external video output (because it was wired to the nvidia card).
and yes i know it depends on whether the laptop has a mux or not... it's not always easy to determine, not all reviews go explicitly over this detail...
I just don't want a discrete graphics card. Not an nVidia one for sure.
i have strongs doubts the doxygen webpages i spend time on will get any faster with an nvidia 3090.
They actually feel similar to Firefox running on Linux on an i5-6500's IGP.
It probably depends on the websites you visit. The only thing better on the newer ones is video decoding for YouTube or the like. I usually actively avoid websites with animations and other similar stuff as I find them very unpleasant to use. But I've noticed those typically tend to use the CPU, even on a Zen3 or Intel 11th gen IGP, even on Windows + Edge.
The old unibody MacBook Pro comes to mind as an example of this.
Traditionally yes; the machine is designed to only run the dGPU under heavy load, not all the time, the BIOS setting just disables doing the thing that you did in the "heavy load" condition. I guess there might be machines it wouldn't work on but I've never known any.
I have an RX5600 and an AMD Zen3 with integrated GPU (in separate computers). Both work perfectly under Linux with AMD's in-tree drivers.
I don't have a multi-GPU PC, so I don't know what the switching situation is.
On the Framework 16, I wouldn't be surprised if someone makes that: it looks like the design would accommodate a left/right-hand side narrow input module with the keyboard still centred.
On keyboard layouts like Framework's, I still have trouble trying to touch-type the right control key with my ring finger-nail. I prefer a "75% keyboard" with an navigation/editing column which provides room for the right ctrl key further right. (And full size arrow keys. And makes shift-End easy for selecting rest of line.)
Maybe we can hope a Kickstarter keyboard maker will make a Framework version in a few years. Main limitation is the thinness, only 3.7mm.
I've had similar cursor problems with those keyboards that have page-up and page-dn next to the up-arrow key. Just a little finger displacement when trying to left-arrow and the cursor is suddenly inserting my subsequent typed text far from the intended location.
And my hands are a little big. Not crazy big though.
If this is anything to go off of, then it seems like they are going to support embedding full QMK enabled input devices. Extremely exciting!
The interface for input modules to the rest of the computer is "just" USB 2.0, so you could put just about anything there that'll fit in one of the input module form factors. Although high resolution, high frame rate touchscreens are probably out because of bandwidth constraints (since it is only USB 2.0).
I wonder if we'll see a Kickstarter for it soon...
Unless one can come up with a way to produce a custom keyboard with the same height restriction, using something like a Thinkpad keyboard which is considerably taller would essentially require a custom taller chassis.
Not impossible, but relatively hard I would say.
Though my Thinkpad can still handle another 5 years of wanton abuse I'm sure, so I have all the time in the world.
Man my next laptop is gonna be awesome. Literally everything I want in a laptop, which has never existed until now(well, soon).
Looking forward to purchasing the 16 regardless though. Keep up the amazing work / mission with Framework. Fantastic work to the team.
(Preferably a proper version with thumb clusters.)
At least for me, that would make this laptop a must-buy.
Honestly, I see the biggest problem that people have entering the ortholinear/ergonomic keyboard market is cost: the obvious choice is to target mechanical keyswitches, but you quickly realize your consumer base would rather pay more (in parts cost and construction effort) to get exactlty what they want.
But what if you went as cheap as possible instead? What if you made flat rubber dome boards that cost next to nothing to manufacture; and sold them for $40 a pop? You would probably broaden your target audience by a factor of 100, and have huge profit margins.
Doing that for an integrated laptop keyboard would be even easier to sell. I would gladly use shitty rubber dome switches if it meant having a split ortho layout that I didn't have to lug around separately from my laptop. I would buy that yesterday.
That being said, the company was acquired by Apple and turned into the iPhone and iPod Touch. The keyboards were effectively museum pieces at that point. I wonder if the patent situation would allow for third parties to revisit the design. It'd be interesting to see what could be done with by borrowing ideas from the programmable keyboard community and mobile touch hardware.
[1] https://en.wikipedia.org/wiki/FingerWorks
They are checking all my boxes now. Outstanding.
It’s often the first thing people upgrade on a desktop (it’s much more impactful than a RAM and CPU upgrade over the same timeframe), but is rarely an option for laptop users.
I'm still happy with my 11th gen framework. Anybody who's looking into upgrading please consider re-using your mainboard. They have 3D printing schematics you can use to turn it into a mini desktop PC!: https://github.com/FrameworkComputer/Mainboard
tl;dr: A 16" laptop with modular dedicated graphics (PCIe x8 either v4 or v5 with fans on the module), expansion bay cards (dual m.s SSDs), a new 'input system' with multiple keyboards in multiple languages and optional RGB lighting as well as support for numpads and secondary displays. Target launch is 'late 2023'.
Personally, looking at the 16...it's just a 'You were supposed to be the chosen one.' gif.
Damn near everything is proprietary, even if they were to open source it, they have basically created their own new standards for everything from the GPU to the new RGB modules that seem like insane distractions. Also those trackpad spacers are a solution but yikes that looks like trash.
All I see is SKU overload and inventory problems all while they are distracted with solving RGB for some wild reason.
Reminded me of https://xkcd.com/927/
If we’re going from 2 open standards to 3 that seems unnecessary. If we’re going from 0 open standards to 1 that seems like a huge step forward.
Like, this isn't a standard at all, it's just open for Framework to gain from. It reminds me of Google's Project Ara but for the wrong reasons.
https://www.theverge.com/2023/3/23/23652967/framework-laptop...
NVidia hasn't made MXM compatible GPUs for at least 5 years.
This includes cooling, so a small GPU could have smaller heatsinks and a beefy GPU could have huge heatsinks.
And if you don't want the GPU and the associated power draw plus the size and weight of the heatsinks, you can swap it for another module if you don't want to game on the go.
As they said in their presentation, it's not just for GPUs. You could have a supplemental battery. They will offer a storage expansion that gives you more M.2 slots. Hypothetically you could even just make a module that lets you fit a desktop GPU on the back, if you provided external power.
[1] https://frame.work/products/cooler-master-mainboard-case
https://cellsaviors.com/blog/how-to-salvage-18650-cells
These days you can go even easier with fully installed plug-and-play Arcade1Up control panel replacements that are literally drop-in - MDF, sticks, buttons, encoders, and audio/display driver card are all built in. You just pull the stock control deck out, drop the $150 Intec deck in, and connect whatever PC/console/FPGA into the video in and USB controller outputs that it provides. You can build a cab around one out of plywood or just grab a random Arcade1Up on sale somewhere.
[0] https://frame.work/products/cooler-master-mainboard-case [1] https://github.com/FrameworkComputer/Mainboard
I also looked at TB4 disk enclosures, but the bandwidth was overkill for spinning rust hard disks. There's a lot you can do with 4 USB-C/TB4 ports. I also have a Coral USB stick connected for frigate.video inference and a 2.5Gbps USB NIC (waiting for better thunderbolt 10Gbps NICs to become available). One fun convergence is when I want a keyboard / display / mouse I can plug the mainboard server into the same thunderbolt dock I use for my laptop peripherals.
[1] https://sabrent.com/products/ds-sc5b
[2] https://github.com/FrameworkComputer/Mainboard/tree/main/Mec...
Don't have time to check now with the site being so slow but will do later. Good to see the amount of attention this receives though.
New 16" Model:
https://frame.work/blog/introducing-the-framework-laptop-16
New 13" Models with Intel 13th Gen and AMD 7000 series
https://frame.work/blog/framework-laptop-13-with-13th-gen-in...
But the linked Github repos for the (GPU) expansion bay and the input module are still private/dead. :/
https://github.com/FrameworkComputer/ExpansionBay
https://github.com/FrameworkComputer/InputModules
Is there an upgrade path from the current Framework to the 16 inch chassis or will you have to buy an entirely new laptop?
Does anybody know why they do this country by country roll out? I'd be perfectly happy to order a US layout laptop but they don't allow me to place an order.
It's probably harder being a US company selling into each EU country than it is for an EU-based company to sell to the rest of the EU and also the US, too?
The alternative would be not selling directly to consumer, using some intermediary like Amazon or whatever.
Do we have an option to add a keyboard with a Trackpoint?
But I guess theoretically someone could design one (╯°□°）╯ ┻┻ they are releasing hardware and software interface docs on github.
Framework gives its modular laptops 13th-gen Intel CPUs and (finally) an AMD option
https://arstechnica.com/gadgets/2023/03/framework-gives-its-...
Already pre-ordered the AMD 13 but might hold off if it's available on the larger version.
Might I suggest engaging with QMK early so that we can avoid the usual manufacturer "hey can you merge this, we've shipped already and forgot to raise the PR until now"?
The height limitations of the 16" laptop are too low for low-profile discrete mechanical switches.
Uh... completely unrelated question: have you laid out any plans regarding module backcompatibility on future boards?
(I have no interest in big laptops. The 13 is probably the physically biggest laptop I've owned in a very long time.)
https://github.com/qmk/qmk_firmware/pulls?q=is%3Apr+sort%3Au...
* Is this an "egpu connected by a cable" situation, or something actually integrated into the laptop's body?
* What are realistic possibilities for how much gpu, and how much of a premium would it be over other laptop or desktop gpu's? Is this something you would be making, or third parties?
> The Framework Laptop 16 has an Expansion Bay system that contains both the cooling fans and an interface for high-power, high-speed PCIe peripherals like GPUs. Expansion Bay modules can extend in both thickness and depth to accomodate higher power requirements.
https://github.com/FrameworkComputer/ExpansionBay
Looking into getting the AMD one for Linux! Has the Battery Life issues been fixed?
Congrats on the Launch. This might be Framework Laptop I finally could use as my daily driver for regular dev. work etc.
On the 16 the keyboard is a USB keyboard that could even be used standalone, without the laptop.
Specifically poor battery life (especially during standby when using the USB-A module) and not being able to use the brightness keys unless you disable the sensor?
[1]: https://community.frame.work/t/12th-gen-intel-core-bios-3-06...
The 16 looks ideal, but I do not have the cash to buy another laptop.
Or that you'll partner with someone who does?
Do you plan providing more options for the display assembly in the future (I'm particularly interested into an option with a touchscreen)?
i might have said before that i regret getting the framework, i might think about upgrading my mainboard and battery if it turns out to be working well. key word "maybe"...