China has released the latest processor based on its own unique architecture.

Intel has leaked new details about its 12th generation Core processors, known as Alder Lake, which was first revealed in January 2021. Videocardz has found that at least one of the new CPUs will show a 20% performance increase in single-threaded mode and immediately 100 percent – in multi-threaded.

Intel disclosed this data in its presentation, the slides from which were leaked to the Web. It’s not yet clear which chips it’s comparing Alder Lake to, but according to Meowing, it could be Rocket Lake, the 11th generation Core desktop processors that Intel launched in March 2021.

According to Meowing experts, several factors or their combination can lead to such an impressive increase in productivity. According to them, the key may be to increase the clock frequency, increase the IPC parameter (the number of instructions executed per clock) and the number of cores.

The slide from the presentation shows that the Alder Lake series processor received a total of 16 cores, divided into two clusters. Intel is the first to use such a solution – all of its existing CPUs have only one cluster of cores, and it spied on the idea of ​​\u200b\u200bseparating them from ARM chip manufacturers, where such an arrangement became the norm many years ago.

Intel will fit two clusters of eight different cores in one processor

The main eight cores are high-performance Golden Coves highlighted in gold on the slide. The eight remaining are Gracepoint, and here there is a direct difference from most ARM solutions. In ARM processors, the cores are usually divided into fast and energy efficient, while Gracepoint cannot be attributed to either of these types – in terms of their capabilities, they are in the middle between them.

Appearance and packaging

The appearance of the packaging practically does not differ from the already tested AMD A8-3850. The graphic design of the packaging is as bright and catchy as in the case of the top model of the A8 family. Of course, on the front side, the developers focus on the features of the APU.

On the left side wall there is a transparent window through which a potential buyer can get acquainted with the processor and its marking. As a “bonus”, the developers cite the advantages of the hybrid processor, such as: support for the built-in DirectX 11 graphics core, energy efficiency, small dimensions. Below is a logo with the class designation AMD VISION A6. APUs in this class are designed for advanced multitasking, photo editing, HD video playback, and light gaming.

On the top side there is a familiar sticker with the key characteristics of the proposed processor: model (A6 3650), clock speed (2.60 GHz), cache memory (4 MB), processor socket (FM1), serial number and product code.

On the right side, developers focus on support for AMD Dual Graphics technology. This technology actually implements the Crossfire mode, in which the load is simultaneously distributed to the integrated and discrete graphics cores.

According to the information provided, the manufacturer recommends using the APU in combination with AMD Radeon HD 6670, AMD Radeon HD 6570 or AMD Radeon HD 6450 video cards. It is with their use that the stable operation of the AMD Dual Graphics mode should be ensured with a noticeable effect.

As you can see, the processors continue their development, and the bundle does not fundamentally change. So with the AMD APU A6-3650 in the boxed version of the package you will see a cooling system, instructions for installing the processor, which also contains information about warranty obligations, and a branded sticker for the PC case indicating the model range.

The complete cooling system is marked Z7UH01R101 and is no longer a novelty for us, because. we encountered it while getting acquainted with the AMD A8-3850, where, in fact, you can read about it in more detail. In short, it can be described as a budget solution that will provide the temperature regime required for a hybrid processor.

but without a margin of efficiency. As for the noise background it creates, then at the beginning of its use it will be quite noticeable. Therefore, if the system unit is placed close to the user over time, it may need to be replaced with a more efficient and quiet cooling system.

It will naturally become the most productive chip among
existing and already being developed in this line. Basic Applications
processor: high performance computing (supercomputing) and systems
data storage.

MCST developing Elbrus, Alexander
Kim. He did not give other characteristics of the future chip.

Another representative could not do this in a conversation with News
MCST Maxim Gorstein. According to him,
in the nearest “road map” for the development of the line in relation to “Elbrus-32C” will be
7 nm is indicated (not 6 nm), but there will be no other specific parameters in it.

As Gorstein notes, some of those planned for release
processors the company already has pre-orders. Part of the circulation will go to
unknown projects. However, to name the approximate ratio of chips of the first and second
category News interlocutor found it difficult.

Note that all Russian developers are the most
advanced processors of small nanometers are still forced to order their release
abroad – usually at the Taiwanese TSMC factory. At the same time, its specificity
work is such that from the order to the arrival of the finished product in Russia, it takes
approximately 6 months. Therefore, taking into account the fact that the MCST calculates in 2021
get the new chips at their disposal, the orders the company will obviously

As for our country, on its territory the necessary
production capacity is not yet available. Creation of factories in Russia capable of
to produce chips with a topology of 28 nm and below (up to 5 nm) is prescribed in
until 2030 – but without any clear timeline.

In the meantime, the most developed Zelenograd
the company “Mikron” has established a serial production of products only according to the standards of 90 nm
and acquired the capacities capable of producing processors on a 65 nm topology
for experimental design development.