2023 is the year of the heterogeneous chip at AMD, if on the desktop we have already seen the configuration of asymmetric chiplets with the high-end Ryzen 7000X3D, the next turning point is for processors for laptops. More precisely with the chip which receives the code name of Phoenix 2 and that it would be already finished and in the last phases before its mass production.
The increase in the cost of manufacturing chips meant that the area of the processor was taken into account when creating new designs, and with this new architecture variants appeared, designed to cope with this new situation. If the concept of heterogeneous cores suggested that it was a mess of Intel, the reality is very different and AMD will also have to adopt the same strategy.
Phoenix 2, AMD’s first wearable heterogeneous chip
Lisa Su’s company’s bet on laptops has always been its so-called APUs, monolithic chips and, therefore, in one piece, with a cached version of its latest desktop architecture in terms of CPU and powerful integrated graphics that stand out from the competition. . However, the high cost of wafers leads them to have to opt for heterogeneous chips for laptops.
And no, we’re not talking about moving the Ryzen 7950X3D and 7900X3D to laptops, hopefully, but about using Zen 4 and Zen 4C cores in a future chip. Specifically it would be an engineering sample that would have appeared in the database of some benchmarks, more specifically it would have the serial number 100-000000931-21_N [Family 25 Model 120 Stepping 0].
What are the Zen 4Cs?
To better understand the news, we must explain that it is the Zen 4C cores that Phoenix 2 also assimilates, which are the other processor cores that accompany the Zen 4 in the heterogeneous design that we are talking about. However, there are differences between the two types of cores.
- The amount of L3 cache is half as much in Zen 4C as in Zen 4.
- Architecturally they are identical, unlike Intel’s E-Cores they are not a different core.
- Zen 4C is designed to be denser, its density is much higher, so more cores can fit in less space. However, it can reach much lower clock speeds. It therefore has less performance than a standard Zen 4.
They were originally designed for cloud computing servers, where a large number of clients require a large number of cores, but with no intention of limiting performance or spending too many resources. It’s part of AMD’s methodology of using existing resources to create solutions to new problems. In any case, despite the fact that the approach is the same as the E-Cores, obtaining more cores per zone, there is no difference in performance between the two types of cores within the chip.