- Intel integration of the CPU nuclei
- The molten cores implement the instructions in parallel before rearrangement to improve performance
- This approach targets a higher efficiency of individual links without expanding the basic size
Inel a Patent What he calls SUPER CORES, which is a technique that merges two or more nuclei the CPU in the apparent “Super Core”.
To the operating system, the pictorial cores appear as one unit, but the instructions are divided and implemented in parallel before rearrangement, with the aim of improving single performance without the high costs of building larger processors.
This approach is similar to the concepts of “excessive orthodontic” from the Pentium 4 era, indicating that Intel reconsidates previous experiences with modern improvements.
Balance of efficiency and size
The idea behind this approach is to improve the performance of a single strand by avoiding the higher energy requirements associated with the clock speed or wider cores.
Instead, the Intel design distributes work burdens through multiple cores through common and synchronous memory units.
If this mechanism succeeds, the company expects performance gains for every Watt, allowing processors to switch between regular and basic conditions.
Observers compared the idea of Intel The old AMD multi -thread collection, although the styles vary.
The AMD is divided into units, while the Intel suggestion combines full cores under the control of the software.
Some also link the patent to the Royal Core project that was canceled from Intel, which has chased high instructions for each hour but has become impractical for manufacturing.
By reviving such strategies, it appears that Intel is looking for alternatives to the expansion of brute force design.
However, the lack of measured data makes it impossible to know if this can compete with the fastest designs of the central processing unit in the market.
The patent describes a small synchronization unit within each nucleus, supported by a reserved memory area called the hole address.
This handle is registration, demand, and data flow to ensure the safety of the instructions.
On the side of the program, dividing translators or binary devices into controlled blocks while entering the flow control orders.
The operating systems must then decide when the work burden benefits from the Super Core mode, a condition that may hold scheduling and consensus.
Without widespread support from both devices and programs, design risks to become an unused feature.
Intel documents do not estimate clear performance gains, which indicates only that two narrower cores may approach the ability of a wider essence under certain circumstances.
Technology can be interested in exploring specialized work burdens, including scenarios in which the CPU may seek to improve efficiency in single tasks.
However, for public computing, the lack of proven criteria leaves the promise that is uncertain, and whether this actually creates the best central treatment unit to demand the burden of work is still an open question.
via Thoms devices