Intel vs AMD: Top Linux Laptop Performance Comparison 2024
ฝัง
- เผยแพร่เมื่อ 1 พ.ค. 2024
- We take the System76 Lemur Pro (Intel Core Ultra 7 155U) and the Framework 13 (AMD 7840u) and run Fedora Workstation 40 @ QHD testing web browsing (Firefox this time!), gaming, and compiling. What team are you on?
The best book (experts only, don't start with this) on performance: Brendan Gregg's System Performance, 2nd Edition: amzn.to/3Qs5Ees
Beginner wanting to learn how to navigate the Linux CLI? Check out:
The Linux Command Line (NoStarch Press): amzn.to/4bhWJnV
Want email that doesn't spy on you? Check out Fastmail (way better than... other providers) and use this code for 10% off: join.fastmail.com/2972c0bf - วิทยาศาสตร์และเทคโนโลยี
You mentioned the E and P cores but one thing you could have highlighted is that this is the U category of Intel which is rated for 15W. The AMD says 28W I believe so it's basically battery life vs performance tradeoff and Lemur is design for maximum battery life.
I would have chooses the newest Tuxedo Pulse 14 Gen 4, it has the newest AMD 8840 chip, and none of the Framework BIOS problems...
Can you compile C++, Rust and Java? If so, then I give you a link to a blog where you can get sources, compile and run and then see results. It should help to decide about performance.
Really interesting video - many thanks 🙏
AMD's APU is a no-brainer for gaming
not a great comparison... the AMD laptop has faster RAM and more of it, as well as better cooling
Yeah I compared these as two of the premier Linux compatible laptops at about the same price points. System76 is about to start shipping the Darter Pro (updated for 2024) and that might be a big difference
Great video.but i have a question.If i buy a laptop with linux preinstalled , is it better than using ubuntu Via WSL.
Its usually about goals - freedom, customization and efficiency are big factors in someone choosing Linux. If all of your programs, programming languages etc work on Windows and you enjoy that, then WSL works pretty well with VSCode integration.
The comparison isn't completely fair because a 16" laptop can have a much bigger cooling system then a 13" one, and therefore get better performance, but AMD still won.
Yes but a 13 inch laptop has less screen space which means less computational workload due to less possible pixel locations. and less screen space means less stress on the gpu, which means less heat per watt. Even then, this comparison boils down to drivers and hardware capabilities. What I would like to see is the comparison between AMD's official Arch based Linux kernels with Amd hardware vs intel's clear Linux with intel hardware. Now that would be interesting.
@@josephlh1690 We are both wrong here, I thought the intel laptop is 16" but it's actualy 14,1" but the framework laptop has a higher resolution despite the smaller display (framework= 13.5" and 1504*2256 pixels) (lemur pro= 14.1" and 1200*1920 pixels).
@@erk_0483 Hmm. If the smaller display has a higher max resolution then that could certainly skew results.
Yeah these were both around the same size, same price, and advertise Linux support (for this channel its a good fit). I just saw Hardware Canucks posted a new video called "Intel vs AMD Laptops in 2024" and they have a bunch of great benchmarks (not Linux focused obviously)
For such a direct comparison you have many things going wrong. Somehow my wifey's notebook, Zenbook 14 on 155H has much better results, myself I am on EliteBook 7840U, both on 32GB. You should, for starters, mention what kernel you are running, with which governors/p-states, which distro with which X-server. Your mention of "balanced" does not mention balanced for what specifically? 7840U is hands down better on performance/power(W), but doesn't mean will give you better battery life (again depending on which notebook and battery capacity).
He said that it's Fedora 40 in the description. For both of them.
The performance of that Ultra 7 155U seems very wacky, the performance is the worser/same as my i3 1220P.
If i look on Notebookcheck that Ultra 7 should get about 9400 in geekbench multi.
I wonder if the laptop you use not be able to cool it properly or uses lower TDP limits.
Possibly down to the Intel having unequal capacity RAM on each channel 8GB+48GB and also clocked lower 4800 versus 5600.
It's a good point but the Lemur Pro is incredibly light (its the lightest laptop I have ever held) so that would definitely be a factor
I have a few things to pick, though overall I agree with the conclusion:
- laptop comparisons - if they are different design (aka, not the same brand, and same chassis, and motherboard design, same cooler design etc) make things very unfair, when the context is "Intel vs AMD" and not "Laptop A vs Laptop B"
- one consequence of the above - how do you know that the AMD one consumes less power ? 7840U has 28W as default TDP, while Intel one has 15W. While on this, power draw and TDP are not the same thing. You can still be at 30W TDP while drawing 60W, if you, say, draw that 60W only for several seconds. Regardless, you can see quite some difference even on the same type of CPU because of how the manufacturer configured the laptop & BIOS and the cooling. This, unfortunately, kills the ability to make fair, correct statements about performance and battery life (not just for this review, but in general). Because there are absolutely bad designs in the wild, where one can be both hotter, and with a worse performance, even worse battery life, all at the same time! Using the same chip. And not by 1-5%!
- it was mentioned, but I want to add it too, to detail why it's a downside: the 8+48 GB RAM of the Intel system most likely is at a disadvantage. First, it says that it's at 4800 MHz (which, correctly is 4800 MT/s) which is most likely lower than what the Framework has. And that dual channel... well, I think that DOES matter. Unless both systems have 8+8 GB of RAM, but even then, the Intel one has a soldered 8 GB for which clearly you cannot get an identical stick of 8 GB, that is, with the same timings. And the speedometer and Factorio and compile tests, I think all benefit quite well from better memory performance.
- Factorio as a gaming benchmark ... is a weak benchmark, especially how it was shown here. First of all, it's 2D. So you're basically just comparing CPU performance, not the GPU/iGPU. Secondly, Factorio is a very well optimized game (check out their Friday Factorio Facts for a lot of interesting details on how they developed the game) that's been out for several years already. To run well on new CPUs is kind of a given. Third, a new map, even a wave defense one, is low load. Not to mention that generating a new map, not using the same one, is also not very ... uhm ... scientific.
For this to be more useful, a save game where there is a BIG factory should be used. It would be the same on both systems, eliminating that problem. And it would have real load. When I say big, I mean factories that can automatically generate the endgame "item" - the rocket launch. Also, not sure if it fully matters, using the maximum zoom level should also increase the load, for GPU too (well, more like GPU memory, I think). Lastly, I don't remember them, but factorio has a command or option where you can see performance stats, which should be more useful than the simple FPS. Especially since Factorio (which is more a simulation) uses internally UPS, that is, updates per second, which show the true performance.
Laptop reviews are hard. I recommend to always take the power draw (that is, plot the power draw ever 100-1000ms or so), as with that you can see much better how much is the chip performance vs the laptop design. Haven't checked with the tools in Linux that can offer that. You can also simply have them unplugged and check with a smart plug or a meter. But that also introduces a lot of laptop design variance. One might transform 40W into 30W of usable chip power, while another might be able to only do 25W. And 5W at this level matters a lot.
Great channel, so glad I found it.
Which DESKTOP HARDWARE is more compatible with Linux, and will allow you not only to run it normally, but even optimally?
Test says AMD is absurdly faster, reviewer makes it sound like a tie. Shame.
There is a really good video by Hardware Canucks that just came out (Intel vs AMD in 2024) where they test these much more thoroughly but the conclusions are not far off
lets go intel!
I think the AMD system has a huge benefit.
Basically 100% supported by SteamOS,
and that the Framework laptop has support for a dedicated AMD GPU extension module.
This in my opinion makes the AMD system unbeatable.
💪👍😎
🌱
👾✨💫💨