I am still planning to do some durability tests with the Framework, and one of the things I was planning to do was put it in a backpack and be…not at all nice to the backpack with the laptop inside. Not sure when I’ll get around to it, but hopefully sooner than later. But I’m massively busy with life stuff, so I can’t say for sure when I’ll get to it.
3 Likes
Poor Backpack… Dora’s gonna have some explaining to do…
2 Likes
Need to issue a correction- when I grab the laptop’s bottom chassis from the front corner spaces (next to the trackpad) instead of from behind like CJ did in his review, I can make the deck flex with the creaking sound like in the video with only a slight amount of force. Damn.
I still think the laptop could survive impacts or drops, but I wouldn’t want to sit on it. Hope someone is brave enough to durability test their unit, most likely BigT once he has the time
I did some testing on the amd fw13 and going from 2 to 1 stick of ram only lowers idle power by barely measurable 0.2W. Going from 5600 to 4800 isn’t even measurable. Pretty sure the single stick isn’t for power but for space reasons. They should have just made it a bit thicker and double stacked the ram imo.
They still got a massive market-share, especially in laptops. They did kind of screw up the “I want an i7” factor they had by changing the naming but the average person still doesn’t really know or care about manufacturers. Also amd is barely present in the low end mobile segment where intel dominates with the n100 class chips (good enough cpu and igpu + the best hw decoders on the market make for killer general use machines).
6 Likes
Sorry I didn’t meant to nitpick, the 3.7 was intended as a half-joke. You were right about is how long the battery lasts on a charge. To make matters worse(for FL12, better for FL13) The FL13 7840U is more efficient than FL12
Under load, at low load they are pretty close cause intel seems to be better at power-gating unused stuff. I’d also bet the fw12 is a bit more efficient doing video playback thanks to intels extremely good hw decoders but yeah they certainly aren’t beating zen4 or zen5 performance per w under any kind of load XD.
Since they ended up only using a single memory channel and no usb4 I think it may have been a better idea to go all in on the cheap and low power train and stick like an n150/n355 in the 12 instead of going half way with an i3/5 with it’s legs cut off.
As shown, 7840U’s wattage under battery life test is 6.57W, while FL12’s wattage is 6.68W. I think 6~7W is pretty low, and that’s the battery wattage, the CPU wattage is lower than 6W.
Depends on what is going on. Hw accelerated direct 1080p video playback for a bit under 7w is something my t480s from like 2018 can already do (and the fw 13 can too at this point XD).
1 Like
That’s exactly my thought as well! The FL12 would make a lot of sense with an ultra low power processor like the N350 (passively cooled!) if that resulted in double the battery life and a lower price, even at the expense of performance. FW could easily offer this as another mainboard option.
Maybe double digit percent better battery live but certainly not double. Should be quite a lot cheaper (the mainboard itself is probably around the same price but the chip itself should be quite a bit cheaper, then again we don’t know if framework got a killer deal on the chips they got) but still probably more expensive than you’d expect an n150/n355 platform to be.
Easily may be a bit of an exaggeration, it would add a whole other platform to the list of platforms they are already struggling quite hard to maintain.
1 Like
Yeah, I think the 13th U procs make sense when it comes to BIOS maintenance. They just really need to sort out the battery runtime. I wonder if it’s because the BIOS is too boost-happy, and not capping peak power draw to a lower amount.
Something like the N350 would be nice for certain use cases, providing less heat and more battery life. But there are some limitations with the N chips that would be a downer for other use cases.
The N chips are limited to a max of 16GB of RAM. They are also limited to a max of 3 displays (vs. 4) and a max of 4K60 (vs. 4K120 and 8K60) on those external displays.
Sure, those things probably won’t matter to a lot of folks, but it seems like a lot of those sharing their DIY configs so far have gone with more than 16GB of RAM.
Yeah I’d love one with an n series chip tbh would do more than enough for me, but I get why they’ve gone with what they have, will just have to see when it arrives if it’s gonna be suitable for me - I’m very excited
N Series only “officially” support 16GB of memory but I’ve ran 32GB without issue in my N100 and N150 devices.
4 Likes
Framework has it capped at 40 watts, well under its max.
I’m referring to this behaviour. There’s short momentary peaks beyond what the configured TDP is:
1 Like
Replying to myself, adding this, if I’m not mistaken:
For example, the 11th gen boards seem to get their PL1, PL2, PL4 and Psys set from here:
12th and 13th gens:
1 Like
Would be interesting if someone could do some battery live testing with the p-cores disabled and the power limits adjusted to match common n355 setups.
Pretty sure I have seen reports of n100 setups using 32 and 48gb sticks in the wild. And the processor only having a single channel is a lot less of a drawback if the platform uses only one anyway. Same goes for the lack of usb4, unfortunate but also doesn’t matter if the platform isn’t using it anyway.
P3T is an amd thing, I have not looked into how intel works in anywhere near as much detail.
So the 90% of charger + battery logic is also there but I am not sure it behaves the same on intel. I have not seen any complaints about the intel platforms doing battery flipping before.
That other thread started with OP from the 12the gen, me on the 11th gen. It’s doing it (flipping) alright. (i.e. It has a ‘behaviour’…which was observed…as early as 2021, below)
The latest 11th gen BIOS update (3.22) has mention of an attempted reduction of this behaviour by introducing a “5% float”:
Interesting weird that didn’t become a bigger deal.
I really wish someone from framework would commend on what the reasoning behind those extreme boosting numbers is. I didn’t see significant performance losses for going from 179W P3T to 89W but it may be for a workload I have not tested.