The big attraction this year, among other things, is to see the competition between Intel and AMD with its two new CPU bets. Core 13 vs Ryzen 7000, a fight without quarter in a difficult moment of the market where we start from a solid base like the one that Alder Lake left us as architecture. Well, an engineering sample of a Raptor Lake (or Core 13) was seen and its performance tested, where the results are really interesting.
As a good example of engineering, this processor was frequency-limited, which may be a handicap at first compared to the muscle of the market. But since said comparison would not be fair, what has been done is to equalize the two processors, one Alder Lake, the other Raptor Lake, precisely at 3.8 GHz, and the results are evident.
ES1 Q0D8, the first engineering sample of a Core 13, here is its performance
Before we get down to business, let’s see what this showcase of Intel engineering has to show. What we see is a processor 24 cores and 32 threads a 3.8GHz very low voltage 1,146 volts and a consumption of 65 watts. gets 8 Raptor Cove cores and 16 Gracemont cores attached to 36 MB of L3 share.
Interestingly, this is the same sample we were talking about this morning, the supposed i9-13900but that said, something more needs to be clarified about the frequencies, as new captures show that the P-Core yes they reach 3.8GHzwhile the E Core they work between 1GHz and 2GHzwhich will imply a disparity in the supposed equality that one wanted to achieve for the performance tests.
It is also stated that the processor was tested in a ASUS ROG MAXIMUS Z690 EXTREME with 32GB DDR5-5200MHz, where the whole was able to start without problems despite the fact that the BIOS should not detect the processor. For testing, a large amount of software was used, which is summarized in the following table.
Core 12 versus Core 13, Alder Lake versus Raptor Lake, what can AMD do?
Go ahead, single-core performance was oddly affected in all tests, as if Intel wanted to hide it. The BIOS may not adapt the speed of the TVBas it largely depends on this and the assignments and voltages in each SKU.
In any case, going from 16 total cores to 24 total cores at the same frequency gives very striking average gains, since as you can see the average is not less than +20% counting on the loss suffered at ST.
There are several ways to interpret the data, which we believe should be better for what has been said, on a 30% approximately according to our general calculations: Intel increases a 33% the total number of worker threads, so the scalability would actually be 1:1 over Alder Lake, which is less than likely and proof that there is something in store.
On the other hand, AMD is going to have to deal with a performance superior to a core starting with the disadvantage that it already has with Alder Lake and that if all goes well it will be increased by about 20% in total and more than 30% in Multi Core if the calculations and forecasts do not fail us. So can Zen 4 with its Ryzen 7000 match that? And we’re tossing one last question, is that the reason to bring the Ryzen 7000 X3D for that same year? Beware of Intel who can still surprise…