# Framework Desktop Batch 8 Guild **URL:** \ **Category:** Framework Desktop\ **Created:** [August 14, 2025, 11:06pm UTC](https://community.frame.work/t/framework-desktop-batch-8-guild/73898 "2025-08-14T23:06:05Z")\ **Posts on this page:** 1\ **Showing post:** 16
**Author:** ![FW4TeePee](https://avatars.discourse-cdn.com/v4/letter/f/eb8c5e/32.png) [@FW4TeePee](https://community.frame.work/u/FW4TeePee)\ **Post date:** [September 12, 2025, 11:06pm UTC](https://community.frame.work/t/framework-desktop-batch-8-guild/73898/16 "2025-09-12T23:06:36Z")
There are two quality issues in particular which you can read about in these threads: > [@Noisy PSU fan](https://community.frame.work/t/noisy-psu-fan/74751/70): > > Is there any chance everyone who has taken delivery of their Framework Desktop and experienced the PSU fan cycling on/off could share what temperature the computer is being used at? I start to wonder if perhaps it was lab tested at 20C / 68F or something like that, which is how this problem wasn’t discovered already? For me, my computer room is generally around 26C (79F) and if I were to re-purpose it to my homelab, it would be in a room that is typically around 30C (86F). They have room for th… > [@Mainboard's status LEDs and their meaning](https://community.frame.work/t/mainboards-status-leds-and-their-meaning/74678/16): > > What a let down. You - we- have been waiting for this for a long time and then this happens. I would be so disappointed and I’m sure you had big plans for this so called state of the art technology marvel. I am at a crossroads as well - cancel or throw the dice one more time. God be with you and I hope you can replace it with something more reliable. Please keep us informed on the return process - if was painful or seamless. There are many threads talking about performanc with AI, ROCm, and other software: > [@Benchmarks - How do I compare the Framework Desktop in the real world?](https://community.frame.work/t/benchmarks-how-do-i-compare-the-framework-desktop-in-the-real-world/75103/4): > > For general use the Framework Desktop fantastic, you can take a look at Geekbench numbers. Or for more comprehensive testing: [https://www.phoronix.com/review/ryzen-ai-max-395-9950x-9950x3d/11](https://www.phoronix.com/review/ryzen-ai-max-395-9950x-9950x3d/11) - basically it’s neck and neck w/ a desktop 9550X at lower power to boot - really impressive. Whisper performance on RDNA3 is pretty disappointing. It’ll work, but about on par w/ an RTX 3050 in my testing. I have no idea how a Mac compares for Whisper, however. For LLMs based on [llama2-7b q4\_0 llama.cpp …](https://github.com/ggml-org/llama.cpp/discussions/4167) > [@AMD Strix Halo (Ryzen AI Max+ 395) GPU LLM Performance Tests](https://community.frame.work/t/amd-strix-halo-ryzen-ai-max-395-gpu-llm-performance-tests/72521): > > On and off over the past couple months I’ve been doing ML/AI/LLM testing on Strix Halo (Ryzen AI Max+ 395), specifically with the gfx1151 GPU, in my spare time. I was given the go ahead a while back to mention that this has been on pre-production Framework Desktop hardware, but since I only just recently finished this latest set of sweeps, and with release I assume getting pretty close (nope, I don’t know anything, don’t ask, lol), I figure I’d share the most detailed/definitive LLM inference… > [@Ryzen AI Max shows up in open source Stable Diffusion benchmark](https://community.frame.work/t/ryzen-ai-max-shows-up-in-open-source-stable-diffusion-benchmark/74440): > > Not sure if folks follow this site… and this is very individual config dependant so shouldn’t really be taken as an absolute picture of performance … but someone running a Stable Diffusion clone has posted details of an image generation benchmark… and it’s … underwhelming? [SD WebUI Benchmark Data](https://vladmandic.github.io/sd-extension-system-info/pages/benchmark.html) – and search for ‘8060S’ to see the result. \ Has anyone who has received a Framework Desktop run Stable Diffusion? Are you able to add this ext… Design (and firmware) issues relate primarily to eGPU support and a lack of rear access to the PCIe slot: > [@Request: verify dGPU support](https://community.frame.work/t/request-verify-dgpu-support/69392/57): > > Nvidia 5080 running Case Setup Cables tested Non-working 7900 xtx setup. SMBus removal modified riser > [@PCIe Open Ended Slot Request](https://community.frame.work/t/pcie-open-ended-slot-request/65086/75): > > I hope this topic isn’t dead, because I also would be FAR more interested in getting this motherboard if that slot wasn’t closed off… at the moment I’m looking at PCIe 4x 5.0 risers that could open the slot for me… turns out I don’t think those exist yet… you are literally better off using an M.2 → PCIe 4x 5.0 riser which is ridiculous when the board literally has the slot… and no “stock” case easily supports M.2 → PCIe risers… that said, I guess nice there there at least is a solution… even if… And, of course, the handle > [@Where is the handle?](https://community.frame.work/t/where-is-the-handle/74634/18): > > The new portable Framework Laptop 24. These are some of the contributors to my decision The best thing about the Desktop experience so far, given I don’t yet have one, has been the community. People on this forum are amazing! --- _[View the full topic](https://community.frame.work/t/framework-desktop-batch-8-guild/73898)._