Intel's Arrow Lake flagship takes single-threaded performance crown over AMD and Apple, now leads PassMark benchmark

Intel Arrow Lake CPU
(Image credit: Tieba Baidu)

PassMark software has highlighted a new entry in its CPU Mark single-threaded performance rankings. In a Tweet shared this weekend it embedded a chart showing Intel’s upcoming Core Ultra 9 285K processor (an Arrow Lake-S desktop part) is the new single-core leader in the eponymous benchmark. The firm admits that the tested sample “might be a pre-release engineering sample, but results are looking good.”

PassMark’s assessment seems reasonable, as the Core Ultra 9 285K confidently noses ahead of some of the best CPUs available to consumers, like the Intel Core i9-14900KS, and a host of Apple M3 variants. It is even further ahead of chips like the Core i9-13900K and AMD’s best-placed Ryzen 9 9950X. Expand the embedded Tweet to see the full chart.

The above is great news in the single-core stakes for Intel 'team blue' fans, but multi-core results found on PassMark’s website aren’t as stellar. The upcoming Arrow Lake-S CPU has a score of 46,872 in PassMark’s multi-threaded tests. That’s a score that nestles snugly between the Core i9-13900 and the i7-13700K CPUs. For further context, the aforementioned 14900KS scores 62,502 in this benchmark. The new top-end desktop chip looks rather anemic if it intends to step into the Raptor Lake Refresh part’s shoes.

In an interesting coincidence, Twitter’s momomo_us uncovered an HP Omen 35L gaming desktop listing this weekend (Tweet embedded above).  According to the screenshot, it features the aforementioned Intel Core Ultra 9 285K at its heart. The listing confirms and highlights the CPU specs, and boasts that it offers up to 5.70 GHz turbo boost, 36MB of cache, and 24 cores with 24 threads. We assume that the dire PassMark multi-threaded result vs the 14900KS may be largely due to the thread deficit – with the Arrow Lake only having 24 threads, vs the Raptor Lake Refresh with its 24C/32T configuration.

Concerning the poor multi-threaded performance shown here, we also have to bear in mind that this is a sample of one, and it has been flagged as an engineering sample. Furthermore, there may be motherboard / BIOS plus tuning to be done – as well as other tweaks – before Arrow Lake systems are ready for prime time.

Arrow Lake-S isn’t a long way off, though. The new LGA 1851 processors and 800 series motherboards like top-end Z890 products should be with us later this month.

Mark Tyson
News Editor

Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason.

  • Marlin1975
    At what power envelope though.

    Previous Gen also did well on Syn test. But we know how that's working out for all the duds and power usage issues now.
    Reply
  • TheHerald
    Marlin1975 said:
    At what power envelope though.

    Previous Gen also did well on Syn test. But we know how that's working out for all the duds and power usage issues now.
    Previous gen was a lot more efficient than it's competition in ST workloads.

    Reply
  • helper800
    TheHerald said:
    Previous gen was a lot more efficient than it's competition in ST workloads.

    That chart is missing AMD's 9000 series, though it wouldn't make much of a difference. Chiplet designs are less efficient for single core because of the nature of the I/O die and the infinity fabric having to be powered for any core to do work. This adds a set amount of power required to do any work on any core, but its a relatively fixed amount of power. Stock multi-core efficiency is a different story.
    Reply
  • -Fran-
    TheHerald said:
    Previous gen was a lot more efficient than it's competition in ST workloads.

    Good try on spinning it, as always, like this.

    Can you find others where AMD is behind Intel in ST efficiency? I can see the 7800X3D second in the chart and the 13400F first, but we know which one would be the right pick for your enthusiast.

    What about gaming workloads, for example?

    Regards.
    Reply
  • TheHerald
    -Fran- said:
    Good try on spinning it, as always, like this.

    Can you find others where AMD is behind Intel in ST efficiency? I can see the 7800X3D second in the chart and the 13400F first, but we know which one would be the right pick for your enthusiast.

    What about gaming workloads, for example?

    Regards.
    How the heck am I spinning it? I just posted a graph from TPUP, what the hell man?

    How are gaming workloads relevant here? The thread is about ST performance, guy said Intel consumes too much power at those, I provided some evidence to the contrary. Where is the problem?

    Yes, I can find others where AMD is behind. How about this one?


    Can you people stop it? It's just a company. You don't have to defend it against all evidence to the contrary, damn.
    Reply
  • -Fran-
    TheHerald said:
    How the heck am I spinning it? I just posted a graph from TPUP, what the hell man?

    How are gaming workloads relevant here? The thread is about ST performance, guy said Intel consumes too much power at those, I provided some evidence to the contrary. Where is the problem?

    Yes, I can find others where AMD is behind. How about this one?


    Can you people stop it? It's just a company. You don't have to defend it against all evidence to the contrary, damn.
    We (I'm guessing I'm representing the rest of the world) are not defending anything. We're trying to stop you from spreading false claims and misrepresent findings on tests.

    The two tests you've pointed out are not quite good as you may want to beleive. Specially the second, since it's just power with no measure of performance. Unless you want to test stability or something like that?

    In order to stablish whether one or the other is more efficient, you need average tests not single passes of any specific software to prove your bias in a very underhanded way.

    I'll stop here, since it's a waste to continue.

    Regards.
    Reply
  • TheHerald
    -Fran- said:
    I'll stop here, since it's a waste to continue.
    Yeah you better stop it here. It's a total waste to continue, going against established reviews ain't a winning formula for you.
    Reply
  • Lucky_SLS
    We just have to wait for the desktop M4 chips for a new performance king.
    Reply
  • JRStern
    IOW on single core they're all about the same, and on multi there's no chart but probably similar.
    And even the worst of them is plenty fast for 99% of real users.

    How about versus Lunar Lake (recent laptop chip)?
    Reply
  • mac_angel
    I can't help but wonder if it's thermal throttling much like the 13th and 14th gen.
    Reply