Installed cleanly with DDU, the AMD drivers keep on timing out when watching high resolution videos and similar. My FW16 has the second generation AMD RX 7700S dGPU.
I’m in a similar boat with the 7700S dGPU, but I don’t even need to stress it with high res videos or games. I installed the update this morning and 4 hours later, Windows started having a bunch of graphical glitches and I ended up with an invisible window over half my screen where I couldn’t click anything.
I’m also having lots of instability with the new drivers. Is there anything I can capture to help? My screen goes black and then I get the little AMD popup that the driver times out.
We discovered an issue with the install script regarding support for the new NVIDIA 5070 12GB card. We have corrected this with a newly built driver package; the version number remains unchanged at v3.03, but the build date has been updated to 2026_07_01.
You can download the updated package directly from the new link. The internal team will update the official Knowledge Base article with the new links later today.
Could you try the lates AMD Graphic driver, which release on the end of June. Download Link
Currently, we are using the AMD public driver, so you can also provide report by AMD bug tool.
Further, could you provide more detail about the issue. I will contact the AMD graphic team for further debut.
Even using the latest public AMD drivers, eventually the driver crashes and the GPU cuts out. I always use an external monitor plugged in through Display Port, using the Display Port expansion card as an adapter to be able to plug into the dGPU port in the back. There’s nothing obvious causing it, sometimes it happens more when doing GPU intensive stuff, sometimes it’s just browsing the Internet doing normal things like looking up programming documentation. The previous stable driver release did not have this problem.
I think I have an idea of what’s causing the instability in my case. Often I’ll leave my computer on while I go make coffee or lunch or some other longer task. Windows 11 is set to lock and turn off the screen. For whatever reason, the external monitor doesn’t like this and turns itself off and on, and I think the GPU driver really doesn’t like this. I’ll have to do some more testing to make sure I have an actual repro case for the driver instability, but I’m super busy the next few weeks so I can’t promise I’ll be able to give good info soon.
I’m actually running 26.6.4 which I believe is the latest. I have 2 monitors connected via a Lenovo Thunderbolt 4 Universal dock. I don’t think the v-sync option will help since my monitors don’t support v-sync.
I haven’t pinpointed what causes it but at least once a day everything will freeze for 10-30 seconds. I’m going to try updating to 4.05 BIOS just in case that helps. It just started happening when I updated to this driver bundle so I think it has to do with the AMD driver being updated. But around that time I did also switch to using the official Framework 180w charger.
I’m also getting BSODs once a week. Is it helpful for me to send you the minidump from the one today?
Back from my trip away from my external monitors. Indeed, running stand-alone seems to be perfectly stable. I’m gonna make an effort to turn off my monitor when I lock my computer/walk away from it for a long time to see if that prevents the driver instability. Today, I left the computer locked while I made myself lunch and came back to the usual instability symptoms; image freezing then cutting out, then coming back and periodically repeating the pattern until I restarted the PC. Will report back if turning the monitor off helps.
I’m getting GPU crashes (screen goes blank then comes back after a few seconds) multiple times an hour how. It’s extremely disruptive. The BSODs have stopped though. It seems to happen when I’m browsing the web with Chrome. I can’t even use Zoom anymore because the GPU driver crashes every time I join a call and disconnects me. I’ve reported the issue to AMD a dozen times now. Is there anything else I can do? The driver was recently updated to 26.7.1 and it’s still happening.
After experimenting, it does seem like turning off the external monitor plugged into the dGPU seems to be an effective work-around. But it’s very much not ideal. I hope Framework are listening and looking into this.