AMD Ryzen AI 300 Series Dominates Intel Core Ultra 7 Lunar Lake Performance For Linux Developers & Creators

[Michael]

Michael writes, "The Core Ultra 7 256V was consuming less power than the other laptop CPUs which is nice if you are doing a lot of code compilation on battery". But I read in a review of Lunar Lake on another website that the processor's performance drops considerably when running on battery.

Michael, do you test these laptops on wall power or battery power?

All the laptop benchmarks are on AC power unless otherwise noted explicitly (such as articles comparing battery vs. AC power or similar).

[Anon'ym']

Kernel compilation is IO bound more than CPU.
You need to test it at least with same number of threads fir all platforms.

This tests looks stupid.

[Party0445]

Kernel compilation is IO bound more than CPU.
You need to test it at least with same number of threads fir all platforms.

This tests looks stupid.

If I’m looking to buy either chip why would I want to look at benchmarks that artificially restricts one processor but not the other? Seems to defeat the purpose of benchmarking - comparing products.

[Michael]

In the future, it'd be cool to include test results from an Apple MacBook running Asahi Linux. That might shed some additional light on if the battery life angle Intel seems to be taking is competitive with ARM CPUs.

Unfortunately not really applicable when the Asahi Linux support tends to lag many months behind the latest upstream Apple hardware that many readers would be interested in (i.e. latest vs. latest). And beyond that really don't have the budget to buy Apple hardware each generation for Linux benchmarking.

[blackshard]

Kernel compilation is IO bound more than CPU.
You need to test it at least with same number of threads fir all platforms.

This tests looks stupid.

Hmmm, the IO in compilation is more a side effect rather than the norm. Also, with modern ssd devices, then IO is greatly reduced and you see in the benchmarks that more cores means faster compilation, so it is definitely a cpu-bound task.

[avis]

You are ignorant:

Funny you start by being toxic and offensive, then jump to arguing with me while heavily distorting facts to fit your "I buy laptops for heavy MT tasks, so should other people, thus AMD rules, Intel sucks, and I couldn't care less about battery life".

These are two wildly different CPUs which should have never been directly compared, period. Wait for ARL-H with a comparable number of at least cores.

Luckily you're in a vocal minority though it's alarming that so many people have liked your post. I guess the "AMD being an underdog" mantra is still strong in people's minds.

Thankfully the tech press, not Phoronix, has praised Lunar Lake for all the things that are irrelevant for an average Linux fan who cares not about a laptop being an actual laptop. Normal average people meanwhile buy laptops for mobility.

[oleid]

This new architecture and its fabrication process are promising, the efficiency graphs show a great leap and the gap against AMD is greatly reduced.

After all its the same fabrication process AMD uses: whatever TSMC provides.

[randomfoo]

I recently bought a Lunar Lake laptop myself (MSI Prestige 13 AI+ EVO w/ an Intel Core Ultra 7 258V - boy that naming is awful) and these performance results don't surprise me that much - LNL has 4 P-Cores locked at 17W. But, I mainly got it because it was a great ultralight config at <1kg of weight with a 75Wh battery. My powertop and powerstat testing has whole laptop idling as low as 2.3W and under light use (text editing, browsing) in GNOME it seems to hang around 5-8W, which is not bad.

For me a bunch of stuff (GPU, WiFi) didn't work well w/ 6.11, so I had to go to 6.12rc1 mainline - I do think the Linux support overall is undercooked overall. Also, the suspend on my laptop is wonky. When it does work, it still burns about 0.9% battery/h (almost 30% battery/day), doesn't seem to ever get to PC10 and running S0ixSelftestTool, only ever gets to S0i2.1. That isn't a dealbreaker, assuming suspend-then-hibernate can work, but it also seems to have intermittent RCU timeouts on resume that causes the laptop to immediately go back to suspend and never wake up (not great). Just normal Linux laptop things I guess, but hopefully some kernel updates can iron that out.

I haven't done much testing yet, but on the performance front, one thing that was surprising to me was that even with optimal settings, on CPUE the llama.cpp tg128 results for Llama2 7B Q4_0 was only 9.67 tokens/s. This is significantly lower than on my 7940HS minipc which gets 14.42 tokens/s on CPU. Token generation should be mostly memory bandwidth bound, and Lunar Lake has 128-bit LPDDR5X-8533 vs my 7940HS's 128-bit DDR5-5600 so I would have expected the laptop to do a fair bit better (I ran bandwidth as a sanity check and past 2.5MB of sequential reads, MBW drops off a cliff for some reason.

Anyway, I think LNL hits the spot for the its target market. For people looking for workstations I think ARL-H vs Strix (or better yet, Strix Halo) will be where the fun is.

[oleid]

These are two wildly different CPUs which should have never been directly compared, period. Wait for ARL-H with a comparable number of at least cores.

Nevertheless, the test is interesting if you consider it weighed via power. One would have assumed that the CPU with fewer cores and smaller TPD limit would beat the Ryzen in efficiency. But apparently not.

I guess the "AMD being an underdog" mantra is still strong in people's minds.

Given the market share it still is.

Thankfully the tech press, not Phoronix, has praised Lunar Lake for all the things that are irrelevant for an average Linux fan who cares not about a laptop being an actual laptop. Normal average people meanwhile buy laptops for mobility.

I just bought Yoga Pro 7 from Lenovo with Ryzen AI 9 365 for mobility and on-the-train coding. But certainly, if you want a laptop for casual browsing, Teams calls and maybe watching videos, the Intel part or the Snapdragon is probably the better choice. While the battery time of the Ryzen part is lower, is more than enough for the usual rides I do and thus I prefer multicore performance.

[blackshard]

Funny you start by being toxic and offensive, then jump to arguing with me while heavily distorting facts to fit your "I buy laptops for heavy MT tasks, so should other people, thus AMD rules, Intel sucks, and I couldn't care less about battery life".

You were ignorant because you ignored some facts. And this is a fact too.
Other words from you in this quote are blablabla that I have never said

These are two wildly different CPUs which should have never been directly compared, period. Wait for ARL-H with a comparable number of at least cores.

Luckily you're in a vocal minority though it's alarming that so many people have liked your post. I guess the "AMD being an underdog" mantra is still strong in people's minds.

You can read that just that in the post after: https://www.phoronix.com/forums/foru...54#post1496054

Comparison against raw performance has to be taken with a grain of salt, comparison against overall efficiency instead tells a sensible and interesting story.

Thankfully the tech press, not Phoronix, has praised Lunar Lake for all the things that are irrelevant for an average Linux fan who cares not about a laptop being an actual laptop. Normal average people meanwhile buy laptops for mobility.

You can go post on the "tech press" forums then