Thanks for the timely response, but you linked the Framework 13 Driver Bundle instead of the Framework 16 we were discussing.
As an added bonus, I just realized that there is an updated driver bundle (released 10/2/24) for the Framework 16! Though, frustratingly, it contains old drivers (compared to what AMD released yesterday with support for AFMF2) instead of only finally adding support for AFMF1.
I believe it’s the same driver bundle for Framework Laptop 13 AMD and Framework Laptop 16, the link you have shared is also correct. And yes, it does not contain what AMD released yesterday.
In the new driver package for Framework 16, the drivers are old, for the GPU they are older than 24.8.1, for the sound the driver is the same as before, as well as others. We are waiting for new drivers from AMD.
I believe the driver bundle was released back in April, we just patched it so that it works with 24H2. We should be releasing a new driver bundle but I don’t have the ETA yet.
I am excited to try it since I like to think of myself as a super-duper user but I JUST draw fancy things using Solidworks. Therefore that is the only reason I upgraded to graphic (amd-software-adrenalin-edition-24.7.1) since my Solidworks 2019/2021/2024 was not working as intended with the first 16 bundle (Framework_Laptop_13_and_16_AMD_Ryzen_7040_driver_bundle_W11__2024_04_02.exe). I also have the RX 7700s expansion so my Solidworks Visualize which currently does not work with 24.8.1 so I will report back after trying this new "driver bundle : Framework Laptop 13 Ryzen 7040 BIOS 3.05 Release and Driver Bundle " I appreciate the response my friend.
The driver Framework supplies has this bug which leads to the laptop spinning a single core constantly after a day of uptime. Restarting every day is a hassle, and 24.8.1 fixes the issue without introducing new ones (mostly).
But I understand that there are additional processes FW goes through to release the driver. Just takes quite a while…
Keep in mind that the driver bundle that Framework includes in the bundle has been thoroughly tested before its release. Bleeding edge driver releases (especially graphics ones) work with most systems though there are known issues and new incompatibilities as a result of the newest driver releases (hence all the subversions 24.x.x etc.)
In short, what Framework packages with the bundle will be the known stable set of drivers that has been tested with their hardware. The “latest driver bundle” is this constant moving target of balancing what works with the newest vendor driver releases.
Updating to newer drivers may get bring new features and may bring on headaches too. Just be aware when updating outside what is suggested/supported.
Those that use their machines for CAD/creativity/design, etc. are best suited with the most stable drivers available for their hardware. It HAS to work or productivity suffers.
If you get a black screen when installing 24.9.1, change the settings in the iGPU BIOS from Auto to Games.
I rolled back from 24.9.1. I had no black screen but the Explorer only showed icons of files and folder but no text.
I wish this worked for me.
My computer already had the iGPU set to Games. I tried switching it to Auto, rebooting, then switching back to Games, and then reinstalling 24.9.1 but ended up with the same issues.
I did end up installing it under Safe Mode, which didn’t run the new drivers immediately saving the first run until after a reboot, and when it rebooted, I got a message that Windows detected an issue with the latest driver for the iGPU (780M), and automatically disabled it. Adrenalin won’t run or offer any of the features it provides until I get the iGPU driver fixed, but it does now run without a black screen… so a minor improvement.
Still a major disruption for my weekend, but at least my computer isn’t bricked.
I understand the value of known stable drivers, and especially validated drivers. But when working in software development, gaming, or other leading edge creative industries, it HAS to work, while also not being months behind competitors/the general public. So stable drivers (24.9.1 is not a pre-release or beta version) are generally used as the norm.
The fact that Framework only validates drivers once in a blue moon is an issue, both for users that want or need newly released features, but also for anyone that needs security/stability improvements that often come with new stable releases. As @Destroya said earlier in this thread, Framework hasn’t updated their validated drivers for the Framework 16 since April except for a tiny change to fix a compatibility issue Windows introduced recently.
And the different subversions of AMD’s releases signify when the driver package was built for testing. 24 being 2024, 9 being September, and 1 being the first build scheduled for release that month. The fact that it was built near the end of the month (Windows Device Manager lists the build date for these drivers as 9/26/2024), and released on 10/1/2024 (American date format, sorry rest of the world) means that a lot of folks feel like it’s a month late, being the “September release” ‘finally coming out in October.’ It also means that there can be skipped subversions of their releases if they don’t build any for release in a given month, so the numbers don’t mean much in terms of the actual number of releases they ship, or reflect the large or small changes that version numbers historically would represent. This change is also common in other developers/software releases, such as the version numbers with Ubuntu (often targeting release builds in April and October of a given year, so 24.04 or 24.10 would be the expected major versions they release this year with alliterative code names to help keep versions together).
I had 24.8.1 installed and upgraded from within adrenaline software, 24.9.1 seems to be working fine for me on dGPU and implements AMDs new framegen edition.
I also rolled back to the GPU driver from the Framework driver bundle on my Windows 11 23H2 install.
In 24.8.1, the GPU driver crashed every 2-3 minutes during video playback, coincident with this message in the event log:
Display driver amduw23g stopped responding and has successfully recovered.
In 24.9.1, the GPU driver failed to start at all. The system fell back to the generic unaccelerated display driver. In Device Manager, the GPU device’s status description was something like “The device is disabled because its driver failed to start”. System shutdown hung, requiring a hold-the-power-button turnoff.
Rant: Generally, I ignore vendor-supplied drivers because they almost always are old, contain unwanted bloatware, or have baked-in choices I don’t like. Framework so far seems better than most, but even there, I prefer to install each driver manually rather than running Framework’s installer. I do that so that I can install AMD GPU drivers in “driver only” mode and with telemetry disabled.
In case you didn’t already know: if you use a third-party zip unlocker (I use 7-Zip installed via Ninite), you can use it to “unzip” the Framework Driver bundle, and it will result with a folder containing each of the separate driver installers included in the driver bundle.
As drivers are for both FW13 & 16 (I have 13) i can also confirm major issues with newest Adrenalin SW.
I had Adrenalin installed as well. However, the newest update also breaks my system
Gets stuck when /not shutting down (trying to reboot(ing) / black screen and my cpu fan goes crazy.
After forced shutdown the restart takes long and I’m getting serveral AMD error messages pointing to driver update failure. Only uninstalling amddriveruninstaller fixes all issues but replaces graphics driver with standard Microsoft and multimedia driver non-functional.
I initially assumed that the issue was with the relatively uncommonly used RX7700S dGPU used in the Framework 16 (and an Asus?), but it’s looking increasingly clear that the software issues are more serious for the iGPU than the dGPU. I’m curious if there’s any strong commonalities between the systems that are having issues vs the ones that seem to be relatively problem-free.
I am running
And, as stated above, I have been unable to update to 24.9.1 without the system crashing to a black screen, and the system fan ramping to 100%.
My 2c: I reinstalled my drivers back from 24.9.1 to the latest Framework-supplied drivers after 24.9.1 broke Helldivers 2 for me. I didn’t have any other problems, but the game would crash pretty much as soon as I stepped out of the hellpod. That’s been my go-to for time wasting lately, so it was enough to make me roll back. Unfortunately, the latest Framework-supplied drivers are from before Helldivers even launched, so they come with their own set of bugs. Just not game-crashing ones.
I, too, would love to see some more frequent validated driver updates. 24.8.1 worked fine for me. And why do I have to go hunting for the latest BIOS, too? When I google “Framework 16 drivers” the top hit is the official drivers page and that still says 3.03. I would have had no clue 3.05 existed if not for this thread.
I have looked far and wide and haven’t found a firmware update for 3.05 for the Framework 16. I have seen a beta firmware update for 3.04 for the Framework 16 from July, but it was still in beta due to issues that were found. There is a 3.05 stable firmware for the AMD Framework 13 that was linked above, but it explicitly says:
Please note that this software is for Framework Laptop 13 (AMD Ryzen™ 7040 Series) ONLY!
For other Framework Laptops, please read the article located HERE.
The forum thread for the beta 3.04 firmware for the Framework hasn’t had any official Framework posts since July 31, except for @Destroya, on August 22, linking to a previous post, acknowledging that Framework was aware of (and tracking) the issues plaguing the update.
3.03 is the current version, we had 3.04 as a beta release. Driver package is updated last week (so that it can work with Windows 24H2).
Why not roll back to v24.8.1? As far as I’m aware, you can install any version freely on top of any other one.