After years of dominating the personal computer processor space, Intel has had a rough couple of years, thanks to strong showing from rival AMD. The last few generations of Ryzen chips have earned a reputation for outperforming their Intel counterparts both on multi-core performance and efficiency.

Intel’s answer? A new kind of chip that combines high-performance and high-efficiency CPU cores in a manner similar to ARM’s big.LITTLE processor architecture, allowing Intel to deliver 12th-gen Alder Lake desktop chips with up to 16 cores and 24 threads. Now that the first reviews are starting to roll in, it looks like Intel made the right move… mostly.

According to multiple reviews of the new Intel Core i9-12900K and Core i5-12600K processors, the new chips not only outperform their 11th-gen counterparts, but also come out ahead of AMD’s Ryzen 9 5950X and Ryzen 5 5600X chips in most tests… sometimes way ahead.

One area where AMD still seems to have an edge though? Performance-per-watt. Intel’s processors consume more power while idle than AMD’s, and significantly more power under heavy load.

That’s a little surprising, since part of the point of combining performance and efficiency cores is that a computer can shift resources to the more efficient cores when necessary. But as Ars Technica points out, another advantage is that those efficiency cores take up less physical space, giving Intel more room for its performance cores.

So while the Core i9-12900K has 8 high-performance cores compared with 16 for the AMD Ryzen 9 5950X, Intel’s cores appear to deliver more performance per core.

Intel has held a lead over AMD in single-core performance for a while, and that also continues with these new chips. While multi-core performance is more important for most of the things we ask modern computers to do, single-core performance definitely comes in handy for certain tasks (like rendering complicated audio processing jobs, something where I regularly feel the limits of my laptop’s 9th-gen Core H-series processor when editing podcasts).

The new chips support PCIe connections and DDR5 memory with speeds up to 4800 MHz, but since DDR5 isnt’ widely available yet there’s also support ofr DDR4-3200 memory.

Since I haven’t had a chance to go hands-on with an Alder Lake-P chip yet, I’m basically summarizing the findings of other reviewers. But if you’d like to check out some of those reviews yourself, here are some places to get started:

Support Liliputing

Liliputing's primary sources of revenue are advertising and affiliate links (if you click the "Shop" button at the top of the page and buy something on Amazon, for example, we'll get a small commission).

But there are several ways you can support the site directly even if you're using an ad blocker* and hate online shopping.

Contribute to our Patreon campaign

or...

Contribute via PayPal

* If you are using an ad blocker like uBlock Origin and seeing a pop-up message at the bottom of the screen, we have a guide that may help you disable it.

8 replies on “Alder Lake review roundup: Intel retakes the crown from AMD (for now)”

  1. Power consumption is a huge stepback in my book, they’re supposed to be more efficient as they move forward.

    Quite ironically, Intel cannot make their 12th-gen cores work with less power now that they’ve gone big.little on us?

    Seems like they missed the bus. Again.

    1. I think it’s premature to draw much conclusions yet.

      It could be a case that 8-Big-cores is beating 16-AMD-cores, albeit using 40% more power. In which case, you would conclude Intel is more efficient than AMD (+110% performance, +40% power).

      It could just be the advantage of DDR5 vs DDR4.

      It could just be Microsoft’s crappy software in Windows 11.

      So it could be that Intel hasn’t even caught up to Ryzen-5000… and they might be in for a hurting with the upcoming Ryzen-6000.

      It’s going to take us a month or two, to get a clear picture (drivers, memory, etc etc). I would look at Hardware Unboxed (YouTube/TechSpot) when they do more videos on this, as they are decently thorough, and they are very unbiased.

        1. I know, I read it.
          If you look at it, it’s quite rushed by AnandTech’s usual standards. They’ve drawn some conclusions which are a bit premature, but at least they’ve noted it.

          In the next 6 months, how Intel-12th versus AMD-5th compares is likely to be (significantly) different. I know this, because it happened to AMD with Zen1. There is going to be a lot of updates to software/Apps, and there are going to be improvements by Microsoft, and lots of new driver updates. I feel like computing will eventually take a paradigm shift towards Intel’s approach, but that will take years to realise. This might mean Intel gets a headstart though. But the midterm future should probably see AMD with the advantage for regular consumers.

          But AMD’s 5th-gen might still be the winner despite everything, if Intel’s 12th-gen turns out to be mostly a paper launch with the current global chip shortages. We are still uncertain about 2022’s market at this point.

          From what it seems:
          – Intel’s 10nm node is slightly matured (~ TSMC 7nm, still behind AMD’s current +7nm design, and way behind their upcoming +5nm products)
          – Intel’s new E-cores is the biggest improvement.
          – Intel’s P-core is basically a new Very-Big Core architecture.

          The E-cores are still a Big-Core design, so it’s competing against Big-Core designs from ARM (Samsung Mongoose, Qualcomm X1, Apple Vortex) and their new variants. Has some resemblance with Intel’s Medium-Core design (Core-M), but isn’t related to Intel’s Small-Core design (Intel Pentium/Celeron/Atom) like we saw in their Cherry Trail products Atom Z8750. The E-core is almost as fast as Intel’s 6th gen Core-i5, but uses notably less power than those 14nm products like they should. This will be a big win for things like SBC, Handhelds, TV Boxes, Tablet/Ultrabooks to catch up with the industry if Intel is able to take advantage.
          The P-Cores are quite a bit less efficient than their E-cores and AMD’s Big-Cores. However, using the doubled bandwidth from DDR5, they’re able to process a lot of data quickly. This is very software dependant, where some applications see no improvement (hence, regression) whilst others see huge improvement (hence, surpassing AMD slightly). Effectively we are seeing 8-Very-Big-Cores, compete against 16-Big-Cores. Hence why we have very mixed results. The E-Cores sometimes help push Intel over the edge, other times they are a hindrance by sapping power away from them, or confusing the schedulers.
          We arrive at the chicken/egg problem. Will software developers move towards one direction, where the hardware architects will need to follow to show their efforts? Or will the hardware ship first, prompting software developers to adapt? The demands of the market aren’t too clear.
          …so what I’m saying is that I can see the merits for a 10-Very-Big-Core design with 0-Big-Cores. Likewise for a 0-Very-Big-Core design with 16-Big-Cores. Or other permutations in between, such as: 1+12, 2+10, 3+9, 4+8 etc etc. It really does revolve around the OS, the softwares in question, and how much you want to balance between efficiency and performance. Android has been tackling this issue for over 5 years now, and it is still far from being solved (my idea for phones is 0-Big, 3-Medium, 5-Small for optimum efficiency/performance and little-to-no throttling).

    2. Efficiency is only an issue when you’re maxing out the 12900K CPU — i.e. when you’re pushing it to 100% on all cores. I suspect Intel wanted to make sure their flagship could deliver flagship performance on all the popular benchmarks

      Under more typical workloads, like gaming, the 12900K is on a par with the high end Ryzen CPUs for power draw, and the efficiency of the 12700K and 12600K CPUs appears to be as good as Ryzen even for demanding productivity tasks, like rendering. The 12600K especially can match or surpass the more expensive 5800X in performance across the board while maintaining a similar power draw.

      If you’re looking for a solution for long running high performance productivity workloads, like rendering videos, or compiling software 24/7, then stick with the 5950X, but otherwise, going with Alder Lake doesn’t lose you anything in terms of efficiency over the equivalent Ryzen performer.

      The current issue is lack of affordable motherboards to go with Alder Lake, but that’s always an issue with new platforms for the first few weeks as manufacturers target the early adopter crowd. This will also happen when AMD rolls out the AM5 based CPUs next year.

  2. Personally I see energy efficient cores in a 125W desktop setup as being absolutely a waste of silicon. If threads are important, get an entry level Threadripper… if energy efficiency is important get a laptop chipset. Adler Lake is effectively a workstation/gaming product, and energy efficiency is not important for either use cases.

    1. look at this perspective, Efficient core are slightly inferior to skylake but they occupy very small space, instead of having 8 perfomance core you can have 6 perfomance core 8 efficient core in same die size.

      those efficient core are not there for saving energy only, they are small in size hence Intel can include lot of them without wasting die space.

      1. Just to put some more context to the issue.
        Intel had their Pentium Architecture in the late-90s, and in the early-00s they miniaturised it and called it the Celeron processor. Then in the late-10s they miniaturised it again, and called it their Atom processor.

        Their Core-i processor that they introduced in the early-00s had a big architectural shift in the early-10s. They’ve essentially been using that same architecture (eg Intel 2600k vs Intel 9900k) albeit with small/gradual improvements along the way. This Intel-12th gen E-core they are introducing is actually a miniaturisation of that Core-i processor. It is basically the new “Celeron”. And if they continue on this path, they will probably miniaturise it again within the next 5-years, on something like a 3nm node, and that will become the new “Atom” chipset.

        This Intel-12th gen P-core they are introducing is a brand new architecture, and it’s akin to a “Very Big Core” design. We will probably see this architecture evolve in the next iterations, until they plateau on the design, and have to go back to the drawing board.

        So people should NOT hype this Intel-12th gen too much. It is more like the start of the shift, ergo AMD Zen, Intel Core-i, AMD Athlon64, Intel-Pentium, etc etc. We will see MUCH more impressive things shortly, as the software adapts and the hardware gets a tad better, just like it happened to the above transitions.

Comments are closed.