Intel Raptor Cove is the name by which the new architecture to be used by Raptor Lake processorsa generation that, as many of our readers will already know, will succeed Alder Lake, and that should be presented between September and October of this year, although its launch will be carried out in stages, as has already happened with previous generations.
The high-efficiency cores of Intel Raptor Lake processors will keep the Gracemont architecture, which means that only high-performance cores will see improvements at the IPC level. We have already told you all this in our special dedicated to the new Intel Raptor Lake-S processors, but recently new information has come out that has left us very surprised, since it indicates that Raptor Cove and Golden Cove are roughly two almost identical architectures.
What does this mean? Well, very simple, that we can consider Raptor Cove as an improved version of Golden Cove, and that this should translate into a one-digit IPC improvement compared to that. We would be talking, in theory, of an increase in between 5% and 9%but we must bear in mind that to these figures we must also add the increase in performance that Intel will achieve with a rise in working frequencies.
Taking into account that according to the latest leaks the Raptor Lake-S processors could move in values of between 5.5 GHz and up to 6 GHz in turbo modethe total single-wire performance improvement (IPC plus working frequencies) may be around between 15% and 20%, in the best case. On the other hand, at the multithread performance level, the Raptor Lake-S series will mark a substantial improvement thanks to a significant increase in the number of high-efficiency cores, which will go from 8 to 16, that is, double compared to Alder Lake-S .
So what improvements will Raptor Lake-S bring over Alder Lake-S?
- IPC increase in high performance cores between 5% and 9%.
- Increased working frequencies.
- Largest amount of combined cache memory (L2+L3).
- Increased number of cores and threads (24 cores and 32 threads maximum, compared to 16 cores and 24 threads for Alder Lake-S).
- DDR5 memory support at 5,600 MHz.
- Between 30 and 40 percent better performance in multithreading.
If we take all of the above and do a comparative analysis with Zen 4, we see that things will be closer than it seemed in the next confrontation between Intel and AMD in the market for high-performance processors for general consumption. Zen 4 promises an increase in single-wire performance of around 15% against Zen 3, a figure that would be enough to beat Alder Lake-S, but not Raptor Lake-S.
On the other hand, Zen 4 is going to maintain, in theory, the maximum of 16 cores and 32 threads on its top-of-the-range processor, the Ryzen 9 7950X, which means that the Intel Core i9-13900K is going to have more cores (8 high performance and 16 high efficiency), and the same number of threads. Most of its cores are going to be high-efficiency, but by having a greater number of cores, I think it is likely that the Intel chip will end up being the most powerful of the new generation, both in single-thread and multi-thread. The numbers point to it, and quite clearly.