Hey all! Just got my Framework 16 in the mail earlier this week. Had a bit of trouble setting it up, nothing that couldn’t be solved with some help from the forums, and now it’s up and running. I’ve been impressed by the gaming performance on the 7700S, but I can’t help noticing the temperatures on my GPU are nothing like what reviewers reported. Jarrod’sTech and LTT both reported GPU temperatures in the mid-60s under stress tests. On the other hand, MSI Kombustor brings my laptop up to 82C in 20.5C ambient, the GPU sits north of 85C in 3D Mark, and anecdotally I’ve seen it scrape 90C a few times while gaming.
To be clear, I don’t think my gaming performance is affected. I’m still getting exactly what I’d expect out of a 7700S, based on other hardware I have on hand. And it’s not like 85C in a GPU is abnormal for a laptop. Maybe FW lowered the fan curves a little bit in last-minute tinkering to trade some thermal headroom for noise? How does this compare to your GPU temps?
EDIT: After posting, decided to run 3DMark a few times looking for actual performance instead of just thermals. My graphics score was 8532, notably lower than NoteBookCheck’s listed 10294. So that’s odd…
2ND EDIT: I’m beginning to wonder if I got a dud. Hardware Canucks shows the GPU going up to 100W and 2.4GHz, whereas mine seems to hover around 1885MHz, 70W, and 90C under full load. Lifting up the chassis to let the thing breathe seems to bring the clocks up to 1950MHz, but when AMD’s website lists the game clock at 2200MHz… I think I’m going to go ahead and open a support ticket.
Are you benchmarking on battery or AC? And which charger are you using? Is windows power profile on performance? Did you install AMD software and pick performance mode for the gpu?
A/C power using the included 180w charger. I have the charge limit set to 90% in the BIOS, but I can’t imagine that has anything to do with it.
I’ve had it on balanced but switched to performance when I started noticing underperformance. I did a few tests and got identical results between the two power modes.
I have AMD Adrenaline installed. I can’t say which “performance mode” option you’re asking about in that regard, but smart access graphics is enabled and every time I start up a game, the screen does that little hitch before the popup saying it switched from APU to GPU so it’s definitely running on the 7700S.
Ok, thanks for the heads up! I’ll poke around a bit deeper in the software and see if I find anything, but in the meantime, based on the temperature and fan noise I’ve also submitted a support ticket asking if it’s expected performance/what can I do about it.
It does sound like you’re overheating there. I tested 3DMark using a 60w charger last night, so I expected a much lower score, I think I got around 7500ish combined score (don’t recall graphics off the top of my head)
Temps never went over the mid 50’s, while pulling about 50w on GPU. Fans were pretty silent too.
You mentioned something about fan noise, was that saying its really loud or quieter than expected? You’d could check that both fans are running properly. I found a slight assembly flaw on mine where a piece of rubber on the bottom of the fan mounting wasn’t fully cut off and would rub against the fan blades making noise. It was obvious on turning the laptop on though, so probably not your issue. Maybe check both fan cables appear to be fully seated?
If you’re comfortable/familiar with repasting laptops, I might recommend taking the GPU module out and redoing the thermal application, or maybe even just checking the heatsink assembly is fully tightened down. If you’re not, checking the the FM support ppl is probably the best bet.
Good suggestion. Visually inspecting, the fans are definitely both running, and since the laptop sounds like a jet engine while gaming I’m going to go ahead and assume they’re both running full tilt, too. I was initially wondering if they were running slow because of Framework’s fan curve implementation, but I got a decibel meter app on my phone and it seems to be the same loudness that reviews mentioned, around 50dB. So I don’t think that’s it.
My own personal suspicion is that you’re right and something is wrong with the cooling interface. I’ve pasted desktop CPUs before but I’m not sure if I’m comfortable working on the liquid metal they use in these laptops, at least not without direct instruction from support. I’ve got a ticket open with them now, and they’ve first instructed me to try all the usual fixes (clean the fans, reinstall drivers, etc), so we’ll see where they go from here.
If it makes you feel any more comfortable with the liquid metal the iFixIt review shows that it is a phase change liquid metal solution so at room temperature it will be a solid and won’t have the ability to spill anywhere. They did rip their metal pad though…
Edit: I do mean for the future. If support offers a replacement that is definitely the better option.
As FYI, the graphics card thermal compound is still standard thermal paste to my knowledge. It was the CPU that has liquid metal applied. If you wanted, it’s probably pretty simple to disassemble the GPU module to repaste it, but again, if you’ve not worked on laptops in this way before, talking to support is still probably best.
Thanks @FlyingTypeTrainer & @Qyygle for the heads ups! I’m not brave enough to go cracking open components just yet, but it’s good to know for the future.
I’ve been going back and forth with support for a few days now trying to diagnose the issue. They’ve got me trying everything under the sun so it might be awhile before I get a resolution. I’ll keep you all updated.
Hey everyone! Quick update on all this: Framework support really did have me running nearly every test under the sun. In addition to thermal issues I also had some coil whine which they spent some time hunting down, and it was lessened somewhat with a driver update from AMD’s website. (I told them it was gone… and then a day later it started cropping up again after extended use. Sigh.) Anyway, coil whine aside, the system was still overheating. Feel free to sanity check me, but I don’t think my graphics card is supposed to hit 90/100° on-die/hotspot in the first minute of a video game. My GPU, on A/C power, was throttled down to 60W at 1600MHz by the end of the first round! They had me run yet more troubleshooting tests, including some that seemed pretty out-there, like removing the expansion cards or speakers.
In total it’s taken a little over a week of back-and-forth (mainly due to time-zone differences), but they’ve finally put me in touch with the RMA department and we’re working out a swap right now. I saw the Reddit post back in May ‘23 about people abusing their warranty claims… but man, they kinda put me through the wringer for this one. I just got the thing, I swear I didn’t break it. I didn’t even touch the GPU until they told me to! Still, I’m happy overall with how they’ve handled the process. Once they determined my system was defective they immediately shipped me a new one and told me to send the old one back in the same packaging. FedEx says it will arrive tomorrow so hopefully everything goes smoothly from here on out.
TLDR: I think I did just get a lemon, and now have the dubious distinction of being possibly the first Framework 16 warranty claim. Ahh well. Thanks for troubleshooting with me!
Ouch. That’s unfortunate.
Yeah, the GPU even on full load for me is always the cooler component, I’ve never seen mine go past 70C. Getting 90-100 on that module is definitely an RMA issue.
Did they have you send back the whole thing or just the module? Curious how they treat that.
I assumed that they were just going to send me a new module, but they actually sent me an entirely new system, minus the RAM and SSD which I purchased separately, with instructions to send the old one back in the packaging. It arrived this afternoon and I transplanted my own internals + the bezel and expansion cards to the new system. The good news is that since it’s an identical system I didn’t have to reinstall a single thing. I just dropped the drive in, reset my password, and reactivated windows. (pro tip to anyone finding this in the future: deactivate device encryption before attempting something like this because that would almost certainly cause you problems. Also, make sure each system is the same BIOS version.) The bad news is that I put a dBrand skin on the old laptop, and upon asking framework support if I could keep the chassis and swap the internals, their reply was basically, “please don’t.”
The new laptop is MUCH more in-line with what I was expecting. Benchmarks line up with where they should be based on reviews and I saw the GPU clocks hitting 2500MHz in DOOM earlier, which the old system never did. I’ve got the old system packed up and I’ll be sending it back to them tomorrow, so I’m thinking this whole thing is resolved!
Unfortunately not. It was just too sticky! I got about a quarter of the way through peeling it off before the printed layer came apart from the adhesive layer. Actually gave me quite a bit of confidence in dbrand’s longevity, but it meant that my new Framework is rocking the “fresh aluminum” look.
Just adding here for others to compare notes to, but yeah, with the dGPU on a high GPU workload, on Best Performance, my temps on CPU / GPU don’t break 70C, and are usually in the high 60’s. If anything I’d say the cooling solution may be a bit over-aggressive on noise vs. cooling demand, but they make be taking things like device skin temperature into account, which is valid. Either way, seeing anything approaching 90-100C on workloads that aren’t power viruses is definitely an indicator of a problem.
