'This CPU is quad-core, much stronger than those of dual-core'
'You see no, this U frequency 4GHz, your i7 frequency is only 3.2GHz'
This kind of remarks we often hear in some treacherous installed store, but if you choose to believe it is usually the old Pentium or Athlon processor, the core number and frequency does have him It's so high, but in fact we all understand the performance. Many years ago, old products obviously couldn't keep up.
Many small white DIY players have begun to tangled up. Before everyone said that the higher the frequency, the more the number of cores, the better the performance, but how is this 'law' is not established yet. To say this, this sentence is true No problem, but in fact there is a prerequisite. It must be the same product or similar product.
Coffee Lake
For example, if the same is the processor of the Intel Coffee Lake architecture, the architecture is the same. In this case, the number of cores and the operating frequency that affect the performance of the processor are greater. In this case, the CPU performance. The number of cores is positively related to the number of cores. That is, the more cores, the higher the main frequency, and the stronger the performance.
However, when the architecture is different, such as the SNB platform and the current new Coffee Lake products, it is obvious that the new platform has an overwhelming advantage.
What is the main frequency?
What is the CPU clock speed? 'The clock frequency of the CPU core's work (CPU Clock Speed) The clock frequency of the CPU shows the speed of the digital pulse signal oscillation in the CPU. It has no direct relationship with the actual computing power of the CPU.' (interpretation from Baidu Wikipedia
Dashboard shapes that are often used to indicate frequency
The speed of oscillation of the digital pulse signal in the CPU is directly related to the speed of the processor. Under the same architecture, the higher the clock speed, the faster the processor is running. This is why overclocking can improve performance. If the architecture is different, It can be simply understood that the effect of a single job is not the same. When two factors are different, the final performance cannot be judged.
AMD Dragon processor
When selecting a processor, it is possible to assist in deciding what kind of product to select based on the price factor. In general, every increase in the primary frequency will increase the corresponding price, and it will be sufficient to satisfy one's own needs. There is no need to deliberately pursue ultra-high prices. Main frequency, mainstream users choose more mid-range products, enthusiasts consider overclocking and other operations.
What is the core?
'CPU core is the core chip of the CPU, made of single crystal silicon, used to complete all calculations, receive / store commands, process data, etc., is the core of digital processing.' ((Also from Baidu Encyclopedia)
Actually, the processors before the early days were all single-core. Later, after continuously increasing the core frequency and manufacturing process, we found that a single computing core could no longer meet our requirements for performance improvement. Therefore, multi-core co-operative computing was developed. Technology, this is the later multi-core processor, so many years later, now more and more mature multi-core technology.
An eight-core processor architecture
Returning to performance, if the computing power of a single core is certain, then the theoretical performance of multiple cores should be n times the single-core performance (n = the number of cores). It is also necessary to ensure that the single-core performance can be achieved under certain conditions. Founded. For example, the single-core performance of the previous generation of AMD processors was relatively weak. After the sharp upgrade of the Ruilong architecture, the current Ruilong quad-core processors have even ridiculed their previous eight-core products.
What is an architecture?
'CPU architecture is a specification defined by CPU vendors for CPU products that belong to the same family. The main purpose is to distinguish the important markings of different types of CPUs.' (from Baidu Encyclopedia)
Threadripper architecture diagram
This definition is a bit unclear. I will explain it in a popular way. The architecture is a way for designers to connect transistors together and let it run calculations. The desktop-class processor uses the X86 architecture, and the mobile end (mainly mobile phones The Snapdragon processor) uses the ARM architecture. The former is a complex instruction set, while the latter is a simplified version. Each has its own advantages and cannot replace each other at the moment.
Ryzen 7 1800X architecture diagram
At present, Intel and AMD's iterative microarchitectures are all based on optimization, adding new instruction sets, etc. According to Moore's Law, every other generation (now reduced to two generations), the number of integrated transistors should be Doubled, the performance feedback should be double the performance, but after ten years of rapid development has now been slowed. However, a substantial increase still occurs after the interval of one or two generations of products, specifically which generation to promote greater, you need to continue to pay attention to New product evaluation.
For example, the AMD processor is a big upgrade, leading directly to the loss of the previous processor product price, completely eliminated.
What is a bus?
'Bus' is a common communication trunk for transmitting information between various functional parts of a computer. It is a transmission harness composed of wires' (from Baidu Encyclopedia) 礀)
Intel and AMD's processor is linked through the bus
The editorial interpretation, generally we refer to the bus here refers to the CPU can support the highest number of bus lines, PCI-E 3.0 X1 bandwidth as a unit, ordinary i7 generally have 20 or 24, of which 16 supply discrete graphics, The South Bridge needs to occupy some (usually four). If you need to expand other devices, you need to consider enough. However, the graphics card can still perform almost all of the performance without occupying 16 buses.
CPU and South Bridge Expansion Diagram
As long as you are not a multi-graphics player, you generally do not need to consider whether the bus is not enough. If you really want to check, the Intel processor is at ark.intel.com/zh-cn/, and the AMD product is directly at the official website www.amd.com/zh- Hans select model query.
How to see the model?
In most cases, the CPU's model can be directly used to infer its performance level. The author will give an example to illustrate. Of course, a more convenient method is to directly query the CPU performance ladder diagram, determine the performance level at a glance, and whether it is suitable for instant clarity.
Intel's eighth-generation Core processor
The Intel processor uses a naming scheme called 'Core i' + '3/5/7', which corresponds to the model from entry to high end. The first four digits in the rear corresponds to the product generation, and the '8' in '8700K' stands for. The eighth-generation Core processor, the last three for the performance distinction, 4790K than the 4770K performance is higher, the last K is not locked multiplier, which is overclocking.
AMD Ryzen 5 1600 processor
AMD Ruilong series processors are also in the form of 'Ryzen' + '3/5/7'. The products are from low to high, and the first four are also representative of generations. The '1' of '1800X' represents the first generation of sharp. Dragon processor, 800 is the same table name is the performance distinction, 1800X is stronger than the 1700X performance, the last X represents the support of XFR technology, can automatically adjust the frequency according to load conditions, AMD full range of products support overclocking operations.
The new APU model number at the end of the 'G' represents the integration of VEGA's single display core.
