Chapter 257 - Flawless Architecture

Chapter 257: Flawless Architecture
Translator: 
Nyoi-Bo Studio 
 Editor: 
Nyoi-Bo Studio

Mid-December.
The
Household Computer
R&D center at Su Hing City achieved a breakthrough.

Mr. Chen, the first batch of prototypes have been produced. Its operations are very smooth. We have achieved success in both the hardware and software system!
Hu Teng Jiao who was in charge of the project said excitedly in the phone call.

Prototypes are out?

Chen Jin was slightly surprised. The
Household Computer
project that had been going on for two years had finally came up with substantial results.


…
The following day.
To show that they attach great importance to this project, Chen Jin’s real self along with a few executive members drove to Su Hing City and inspected the research center.
He sent his greetings to nearly 2,000 software and hardware developers in the research center.
Seeing their boss in person, the developers were every excited. They shook hands with him, took photos with him, and there was even a female staff that asked for a hug from him.

Great job everyone, you’ve done well.

Chen Jin constantly waved his hand. The scene was filled with acute enthusiasm.
…
Inside the product testing room.
Chen Jin saw two prototypes placed on a long table.


One of them was a household desktop. It had a rectangular CPU and a 19-inch LCD display, not much of a difference with ordinary household computers. The other one was a laptop. Because it was a test prototype, its appearance was also rectangular, and its thickness was comparable to the ThinkPad.
Project leader Hu Teng Jiao did not immediately boot the computers. On contrary, he asked someone to bring in two chips.
The computer CPU chips!
One of them had physical dimensions of 4×4-centimeter cube, half the size of a palm. It was four times the size of regular Intel CPU’s… this was the CPU for the desktop. The other chip was 2.5×2.5-centimeter cube… this was the CPU for the laptop.

These two CPU’s were produced using Zhong Xin International’s 28 nm process technology. It was two to three generations behind Intel’s most advanced 10 nm process technology… however, the production of Intel’s 10nm is severely scarce. As of now, they are mainly still making 14 nm CPU’s.

The difficulty to make computer CPU’s were much higher than making the Soc on mobile phones. The surface area of computer CPU’s was a few times larger. The number of transistors was also multiplied several times and even to as many as ten billion times.
While using the same process technology, the larger the area, the lower the conformity rate, and the higher the cost price. Hence, Intel’s 10 nm CPU only realized mass production on 20X0 after being held back for three to four years.
CPU’s made using Zhong Xin International’s 28 nm process technology was certainly worst. The performance gap can only be compensated by increasing the area of the wafer and the number of transistors.
Fortunately, after Chen Jin helped solve the problem of the lithography lens, the Super high-resolution lithography machine that would soon be launched by Shangwei Microelectrics could be used in manufacturing the 10 nm computer CPU, assuring optimal conformity rate… it could almost tie with Intel’s most advanced CPU.
In addition, the size of the CPU here was several times larger, and it would be completely unnecessary to worry about it having underwhelming performance.


More importantly, the CPU, named as the ‘Azure Generation’ used the more common X86 and MIPS for its physical architecture, which both had a huge competitive advantage.
Engineer Liang Yongpin who was responsible for the CPU architecture introduced


Mr. Chen, the CPU architecture Intel and AMD are using as for now is mostly the X86 architecture which occupy the absolute mainstream in the market.


The X86 architecture which uses a complex instruction set features single line instructions with high speed and a smaller number of instructions. It executes instructions sequentially, making it easier to control.


But the X86 instruction set has only 8 general-purpose registers. In CISC, most of the data is accessed in the memory. The RISC (Reduced Instruction Set) systems on the other hand tend to have very many registers, which slows down the speed of the entire system. In some micro-decoding scenarios, the decoding speed will be very slow and complicated…causing the computer to have a low utilization rate and slow operating speed.


The MIPS architecture with a reduced instruction set is similar to the ARM architecture. Consisting of a large number of core registers, it reduces the amount of data access to the memory and greatly reduces power consumption… with the same performance, the MIPS architecture provides the lowest power consumption per square millimeter of the chip design, and is more flexible and open to use.


The shortcoming of the MIPS architecture is that there are certain limitations in terms of memory and cache support. There are some problems when dealing with high-capacity memory… and this is exactly what the X86 architecture is good at.

Liang Yongpin held the chip in his palm, presented it in front of Chen Jin and said,

This is the ‘Weaver Architecture’ this CPU is using. It takes into account all the advantages of the X86 architecture and the MIPS architecture at the same time. It has fast speed, high performance, fast memory data exchange and high CPU utilization, but its power consumption is comparable to the MIPS architecture.


There are definitely shortcomings that exists in other architectures, but on our Weaver Architecture, all of them are unriddled!




The reason is that the transistor distribution on this CPU is no longer done with the previous matrix arrangement, but a special kind of ‘array’. This ‘array’ can be diversified, allowing those transistors to be combined into computing units and sometimes combined into memory cells… transitioning intelligently to meet a variety of usage.


Perfectly actualizing both high performance and low power.

Chen Jin nodded.
The Weaver Architecture was brought over by him from the Base of Operations, of course he was aware of the strength and the advantages of this architecture.
In short, this architecture was in the forefront of all the mainstream architectures that currently existed on the market. Is was an advanced product beyond a generation.
Not only that it enacts both high performance and low power consumption, it also has a killer-like function. And that was, it can be a pure CPU or a pure GPU. It could complete all the operations that graphics card could do without users having to purchase a graphics card separately.
Commonly, the price of a slightly better CPU was approximately a thousand dollars. But the performance of the built-in integrated graphics would be very poor and would only allow users to play games that required low system configuration.
If a better graphics card dedicated to graphics operations were to be added by the user to play some battle royale’s, its price on the other hand would be at least a thousand dollars and above, or maybe even 2000, 3000, 5000… a graphics card would take up about half of the price of the entire computer. The price of the two items combined would be equivalent to 60-70% that of a high spec computer.
Now, courtesy of the Weaver Architectures ‘Azure’ CPU, people won’t have to buy a high-performance graphics card separately at the price of merely 1500 dollars. How could people not be moved?
Of course, considering that both CPU and GPU would greatly increase power consumption, if someone were to run a game that required extremely high system configurations, not even cold water could cool down the immense heat released by the hundreds of watts of running power. Independent graphics card must still be developed later on.


But as of now, it was perfectly adequate.
Hu Teng Jiao held the CPU the size of half of his palm and said,
Mr. Chen, if this CPU can be made using the 10 nm process technology, its performance in graphics processing would be roughly equivalent to the RTX2080… this would suffice in current times. Most users won’t need to consider adding another graphics card.

Chen Jin nodded. The RTX2080 was already a huge thing priced at over 10 thousand dollars on the market. Maybe only not more than 1% of the world population can afford such a deluxe.
Aside from a pint-sized wealthy population, for more than 99% of regular users, a CPU with the ‘Weaver Architecture’ would make the grade for them.
Of course, good hardware architecture must still be paired with a commensurate operating system and software to unleash its utmost potential.
For example, although there were many shortcomings in the X86 architecture, the win-tel alliance formed with Microsoft had played a decisive role in the widespread application of the X86 architecture that it ended up dominating the absolute mainstream in the market.
Hence, Chen Jin set his sight back on the two computers in front of him.

Turn on the computer. Let me see the operating system’s performance.


Okay.

Two staff members simultaneously pressed the power button.