Intel CEO Bob Swan has released an interesting message in his virtual appearance at Computex, where he advocates that we focus more. on the value that processors offer as a platform that in the classic performance benchmarks, an approach that, curiously, was defended by AMD when it was at its worst, that is, in the pre-Zen stage (Bulldozer-based FX generation).
With this I do not mean that Intel is in a bad moment, nothing further from reality. The chip giant breaks revenue records and has recorded results that any technology would want for itself, but it is true that not being able to keep up with AMD. The Comet Lake-S architecture has the same IPC as the Coffee Lake Refresh, has been able to barely raise the core and wire count to 10 and 20, respectively, and has high consumption and very high working temperatures.
In contrast, the Ryzen 3000 processors, based on Zen 2, are manufactured in a completely new architecture, have reached 16 cores and 32 threads and have perfectly balanced temperatures and consumption, so much so that the Ryzen 9 3950X, which has that configuration of 16 cores and 32 threads, consumes less and is cooler than the Core i9 10900K, which has 10 cores and 20 threads.
Why does Intel want us to forget about benchmarks?
We might think that the CEO of Intel wants us to see the value that a processor offers as a whole, that is, evaluating more aspects in addition to the gross performance, but in reality it is clear that the chip giant wants to forget about the benchmarks because they no longer benefit youRather, they put increasing evidence that they are not being able to keep up, that the 14nm process is exhausted, and that they need to complete the 10nm transition as soon as possible.
Valuing a processor as a platform has important implications, but the truth is that Intel wouldn’t do as well as it thinks it would either. AMD has made the leap to advanced technologies, such as the standard PCIE Gen4For example, it sells solutions with very powerful integrated graphics and very low consumption, and it has processors with an excellent price-performance value on the market.
It is true that Intel has an advantage on some fronts. For example, Cascade Lake XE processors and those based on the Ice Lake architecture have AVX-512 and DL-Boost, but its advantages are limited to professional sector, specifically to neural networks and deep learning. If we approach the sector from a traditional perspective focused on the values of price, consumption, temperatures and performance, AMD is the current winner, without any doubt.
Before finishing I want to take this opportunity to remind you that, according to the latest information, Intel will launch processors at the end of the year Eleventh generation core based on Rocket Lake-S, a transition architecture that will be based on the process of 14 nm +++ and that will involve the adaptation of the Willow Cove nuclei to said process. After Willow Cove will come Golden Cove, and finally Ocean Cove, which will increase the IPC by up to 80% against Skylake.
I don’t know