Intel Alder Lake Desktop Launching: 12th Gen Core Processors Detailed
Intel’s Hybrid Core Design Debuts
While Apple has garnered a lot of attention recently with their M1 (and new M1 Pro and M1 Max) processors, there is another player in the hybrid core processor design space, and it just happens to be the biggest player in the game (and by a wide margin, even with AMD’s market share at a 14-year high).
We are of course talking about Intel, and while they certainly have not dropped their venerable x86 architecture in favor of Arm, as Apple did, they have built a new processor family that uses a mix of high-power cores (P-core) and efficiency cores (E-core) just the same – up to eight of each in the high-end SKUs.
This new approach with Alder Lake represents “Intel’s biggest architectural shift in a decade”, as the slide below illustrates:
We took a high-level look at the new core microarchitectures found in Alder Lake during Intel’s Architecture Day 2021, and the company claims that we will see increased performance in both single and multi-threaded workloads with these new chips.
The Unlocked Alder Lake Processors
Before moving on to Intel’s performance claims, let’s first check out some of the products that will make up Intel’s 12th Gen desktop lineup with a look at the unlocked SKUs:
At the top of the chart we have Intel’s new desktop flagship processor, the Core i9-12900K, which is a 16-core, 24-thread part. This uneven core/thread count is explained by the division of P-core and E-core, as only the performance cores offer Hyper Threading. So, the Intel Core i9-12900K is a 16 (8P+8E) Core part. Just as Intel’s processors now carry lengthy names, the descriptions of their core and thread counts have similarly expanded.
This flagship SKU features the same Turbo Boost Max 3.0 frequency of up to 5.2 GHz that we had with the previous generation, though this chart does not mention Thermal Velocity Boost Frequency (5.3 GHz with the Core i9-11900K). Cache, on the other hand, is up considerably from the 11th Gen flagship, with a total of 14MB of L2 and 30MB of shared L3 (Intel Smart Cache) – up from 4MB L2/16MB L3 with the Core i9-11900K.
Another interesting aspect of these new chips is the way that power is rated, and I am very pleased to report that it is actually realistic now! Rather than presenting a blanket TDP of 125W, which is ignored by virtually every motherboard manufacturer, there is now Processor Base Power and Maximum Turbo Power. With the Core i9-12900K the Base Power is 125W while the Maximum Turbo Power is 241W. Finally Intel is providing a much clearer look at the type of cooler such a CPU should be paired with.
An Early Look at Performance
We don’t have benchmarks to share with you today, but Intel did provide some data based on internal testing to give an idea of how these might perform. I will mention that any comparative Windows 11 testing of Ryzen processors must be performed with a fully patched system to address the performance issues with AMD CPUs on the new OS, and, as these results likely predate the latest patches, take these early results with the usual grain of salt:
Regardless of the AMD results, Intel is showing some pretty impressive gains over the 11th Gen flagship here, and if Intel can take back the gaming crown from AMD with this launch it will be quite the shift after the dominance of AMD’s Ryzen 5000 launch.
New Processors, New Platform
With the move to LGA1700 Intel isn’t just using a slightly larger socket with Alder Lake, as there is a new platform – and new memory standard – as well. So, what are all of the changes?
We are talking about the first platform with DDR5 memory, and the first with PCI Express 5.0. After catching up with PCI Express 4.0 last generation, Intel no longer plays second fiddle to AMD with platform innovations with this launch. Of course we will have to wait for PCIe 5.0 products to actually take advantage of the 2x throughput capabilities, but at least 12th Gen buyers are getting a future-proof I/O platform.
Intel has also doubled bandwidth between chipset and CPU with the move to DMI 4.0, has added PCIe 4.0 lanes to the chipset, and has integrated Wi-Fi 6E as well. More features can be found on this Z690 Chipset slide:
We await independent benchmarks, but the Alder Lake launch seems to have the makings of a return to, at the very least, serious competition in the enthusiast desktop space from Intel – if not an outright leadership position. That would be quite a story after the past couple of years. Stay tuned.
For the first time in a long time I’m looking forward to an Intel review. Plus the platform is really interesting too. Finally Ryan is going have something to shrout about.
Does Alder Lake support DDR4-3000 memory? Or, does it need to be 3200? I haven’t bought memory for about 5 years, and it’s DDR4-3000 that I had been planning to bring over from the old board. If I have to buy new memory, regardless, that changes some things, and I’ll need to take a serious look at DDR5.
Just lower timings and if needed up volts a vit to OC it. My 3200 is just pc2666 OCed some. But you will need to do manual timings then. Once learned it is always easy to get right, and never have incompatible BS again. No, mem profiles are not the answer to everything. You can mix and max all you want if you understand what you’re doing in the bios with mem settings. Timings usually come right off the sticker on the chips.
Are you guys going to discuss the Intel SSD Gen 5 performance video they put out on Twitter?