I’ve seen multiple topics on the forum about battery life. But haven’t really seen one that explains what we can expect as a different with the 3 cpu options. Does the i5 model last longer than the highest end i7 option?
1 Like
If I had to hazard a guess I would say probably not as they are all rated to the same TDP, so power consumption is likely similar
I suspect the difference lies in heat output and performance-not power draw
1 Like
Yes since they all have the same TDP, there shouldn’t be much of a difference.
Counterintuitively, the slower CPUs may be slightly worse. On intensive tasks they will have to stay at higher power/higher clock speeds longer. That might have a slight negative impact on battery life.
3 Likes
Oh derp, hadn’t even thought of checking that. That’s nice to see.
@ctl might be something to add to the configuration page (unless you guys think it clutters it to much). But like a little (i) mouseover for the cpu section that just says they’re all the same tdp and wil probably result in very similar battery life.
Cause I can imagine lots of people picking the i5 because they ‘want better battery life’ than with a ‘powerful i7’. I know that i5 and i7 shouldn’t be seen as performance comparisons, but that’s how many people still (incorrectly) think of it.
They are made from basically the same chip, same core count, while i7 has better average silicon quality and can run at lower voltage. But that will NOT be a noticable difference. Undervolt will make much more difference when it is supported in the future.
@Eduimudzz I’ve seen some reviewers measure differences in terms of like an hour extra battery life depending on choice of SSD
The SK Hynix Gold SSD has made a name for itself for this exact regard-it’s PCIe 3.0 but for the majority of users, they’ll never notice the speed difference but they will notice the battery life gain
@Fraoch you probably don’t know but CPUBenchmark is not a reputable site, I’d suggest going elsewhere for comparisons
@GhostLegion yes, because the pcie 4.0 controller in cpu io all so draws a ton of power when it’s on 4.0
It’s more than that, SK Hynix has a unique board design as well, I encourage you to look at some reviews to find out more, both Wendell from Level 1 Techs and AanandTech did some really good reviews
1 Like
The only time the faster i7 is going to give better performance is in the highest p state and im going to assume the Framework strictly adheres to Intel Turbo boost limits and such so realistically the only time this is really going to come into play is long code compilations and I’d argue that anyone doing that is likely plugged into AC power anyways
But I’d love to see some owners post times of how long it takes to go from start to finish compiling Chromium Source and the battery life at the end (assuming they started at 100) I’d be interested in the result
@RandomUser I understand and agree with you, the race to idle favors an intel i7 over an i5…I just think the effect is minimal
The cooling system is well built to deal with thermal requirements and from another post somewhere on this forum it appears it doesn’t throttle under heavy load
I just don’t see most users spending a large portion of time in the highest power state-where a race to idle matters the most
I could be wrong, I’m not a “pro” user, I’m a student in college with minimal requirements, I understand there will be pro users who buy these machines…and I would guess that they would do their work at a desk or something hooked into AC power where battery life is irrelevant
But yeah, fundamentally I agree that an i7 would in theory be better for battery life, I just want some objective testing to show by how much, because 5 percent would be better but it wouldn’t be anything to write home about
1 Like
That’s how it used to be, and how it has been for a very long time. Intel really mixed up its branding in this generation. The i5-1135G7 is definitely lower performance than the i7 in this scenario though.
It’s important to note here that TDP as it’s advertised by Intel, AMD or NVidia often isn’t related to power consumption or heat generation like it should be. While all these CPUs have the same TDP rating, the i5 option has a significantly lower boost clock, which translates to lower power consumption and less heat generated while in a boost state.
A higher quality i7 chip may be able to use less power while not in boost state, depending on BIOS and hardware implementations, but it will also eagerly boost to higher frequencies. Power and heat don’t scale linearly with frequency, or so I’ve been told, so it may be that it will consume more power when in a low usage state.
Of course, power consumption isn’t everything. Usability is a factor too; if you want your PC to be more responsive under heavier loads, a higher grade CPU is one factor in achieving that.
However, I encourage you to look at power consumption from other components too. For instance, SSDs have a surprising amount of variation in characteristics between models. In fact, controversially, sometimes there is variation on the same model, due to different chips being used interchangeably by manufacturers after launch, but that’s another story.
The WD Black SN750 available in the configuration for this laptop has apparently been seen with some kinda meh idle power consumption, but that may be a trade-off for a quicker wake from idle. It’s also known to be very efficient in other scenarios so it’s a bit of a toss up. If you want a highly efficient alternative, the SK Hynix Gold P31 ranks pretty high on efficiency if I’m not mistaken.
This is true on desktop, it’s my understanding that TDP is actually quite accurate on mobile as OEMs don’t like dealing with wishy-washy numbers when designing the cooling setups on laptops
This is true
Only if the power states are set differently, and I’m pretty sure they aren’t, it would defeat the purpose of race to idle to have idle frequency higher than it strictly needs to be
Yep, I said something similar-they have been measurements run that show a significant difference in run time just by switching to a more efficient SSD by as much as an hour
The fact that Intel removed undervolting support on tiger lake annoys the crap outta me and they gimped the low power states as well, crippling battery life on linux
1 Like
Huh. Well, I was already thinking of switching to the Gold P31 and throwing this drive into my gaming PC. I hadn’t guessed the difference would be this big, however. I guess that’s what you get with such a small, low-powered laptop though.
Yeah, I’m running Linux until I’m ready to install Windows 11. The power consumption is miles better than my 5 year old gaming laptop, but it does lose a significant amount of battery when sleeping overnight or when idling.
It’s for power efficiency alone that I’ll switch to the P31-it’s twice as efficient as any other drive, with a gap that large you WILL see gains, I don’t see how you couldn’t
I’m probably remembering in error or incompletely but yeah, Intel ditched the old method of Power States with Tiger Lake and Linux doesn’t have proper support for the new ones so suspend just kills your battery
I have yet to hear what the heck was wrong with the old method-it just screams to me as a way to screw the Linux user
A post from the System 76 community on Reddit seems to have detailed this in technical terms. If I had to guess, it’s because Intel is pushing for their Evo branding. One of the features of an Evo certified laptop is waking to the lock screen in one second, and then connecting to Wi-Fi within one more second. So I guess Intel couldn’t get that working if S3 sleep was enabled? I don’t have any of the tools required to verify that, of course. Also, wouldn’t hibernation invalidate my hypothesis? There’s definitely something going on here…
@Sam_Murray Like I said, this seems fixed to screw with Linux users and I don’t like it
Although I’ll pay the price for garbage battery life if it means a disabled or neutered ME