On August 16, 1985, not long after Nintendo's ninja arcade platform was launched, the R&D personnel of the arcade hardware department of Xinchuang Electronics disassembled and analyzed it.

Most of its CPU, sound card, and display chip are soldered to the motherboard. Therefore, the potential in terms of hardware upgrades is not as good as that of the Phoenix arcade platform of Xinfeng Electronics.

However, there are still memory and graphics card slots reserved on the ninja arcade motherboard, so there is still some room for upgrades.

The storage medium for the Ninja Arcade platform game program is also an HDD. Although its mechanical hard disk is not a hard disk produced by New Venture Electronics, it is a Toshiba hard disk authorized by New Venture Electronics.

At present, the world's mainstream hard disk manufacturers need to pay patent fees to Xinchuang Electronics. For every 100MB of storage capacity, the royalty is $1. Because this patent fee is not too expensive, most manufacturers around the world quickly accepted such a standard.

However, hard disk manufacturers would not have imagined that the storage capacity would gradually upgrade to GB, or even terabytes, in the future. Although the old hard disk technology patents of Xinchuang Electronics will expire, the patented technology will be updated every few years, and the package of patents will still be charged in the past way. Many years later, hard drive manufacturers felt that something was wrong, and when the hard drive capacity grew to 10GB, a hard drive needed to pay $100 in patent fees to the new start-up electronics company. This made the later hard disk manufacturers, from the beginning, think that Xinchuang Electronics is a good person, and gradually become a vampire.

Of course, in order to avoid the rebellion of global hard disk manufacturers, in the mid-90s, the new venture electronics company renegotiated the issue of patent licensing pricing with hard disk manufacturers, marking from 100MB to $1 to 1GB $1 patent fee, temporarily so that all major manufacturers are satisfied. However, in the TB era, these manufacturers found themselves being cheated again......

At present, Toshiba may have produced 256MB HDDs, but Nintendo's Mario Adventure only uses a 128MB HDD as a storage medium.

Because, the current market price of a 128MB mechanical hard disk has been reduced to $150. The price of the 256M hard drive is more than $300. In order to save costs, Mario Adventure has gradually optimized and streamlined to less than 128MB, even if the data capacity exceeded 128MB at the beginning.

As you can see from the capacity of the hard drive, this game is not on the same level as "Court of Avalon".

But after all, this game is a 3D effect, and while the graphics quality can't be compared to Court of Avalon, it's definitely true 3D!

In this era, it is very rare for a game platform to run 3D games smoothly!

"In terms of CPU performance, memory, graphics card capacity and other indicators, the Ninja is about 10% inferior to our 8-core Phoenix, and about 10% behind the 10-core Phoenix. Its only advantage may be that the 3D display is slightly ahead. After a few days of testing by technicians, "we analyzed 16 of its display chip particles. 12 2D display particles, 4 3D display particles. The capacity of the video memory is only 6MB, which is not as good as our video memory capacity. But in terms of display effect, it is a little stronger than our video memory!"

"Why is that?" Lin Qi was puzzled.

"In our estimation, it's the selected memory particles. Filter out particles with higher frequencies. "To run a large program, of course, you have to consider the capacity of video memory and memory first, and if you don't have enough capacity, you can't carry it at all." And the frequency of the chip is secondary to the indicator, which we didn't pay attention to before. And Nintendo's arcade machine, in terms of memory and video memory capacity, does not surpass us, but whether it is memory particles or video memory particles, its frequency is far more than similar products on the market. This shows that it is widely screened, and the higher frequency of the particles from the same batch is selected to improve performance. The remaining memory particles will be sold to the market!"

In fact, there are also some DIY cattle who think that the frequency of the memory modules bought on the market is not high enough, and they are too lazy to use such low-level means as software overclocking. So, I bought my own memory particles, selected some high-frequency particles, and welded them myself. This kind of cattle person needs to be proficient in semiconductor technology and have strong hands-on ability. Of course, there are many such cattle in the semiconductor industry, and many of the founder teams of chip factories do not have wafer processing capabilities at the beginning of their fortunes, but buy some raw materials and solder them into products by themselves, so as to earn the price difference between raw materials and products.

Later, after the price of memory increased, some people also embarked on the road of no return to make memory sticks by themselves in order to save money and have super hands-on ability. The reason why there is no return is because the success rate of handmade is very low, playing is a heartbeat, playing is gambling, and after a slight mistake, the expensive raw materials are scrapped. If you count the total length, most of the gods who make their own memory are more for entertainment and don't necessarily save money.

In fact, the memory produced later by industry is the same, the same batch of memory particles are unqualified, and their performance indicators are also very different.

For example, the current mainstream memory frequency is 66MHz, but if you carefully select, you can also select a memory particle with a frequency of 100MHZ. IN THIS WAY, THE MEMORY READ AND WRITE SPEED WITH A FREQUENCY OF 66MHZ IS DEFINITELY NOT AS GOOD AS 100MHZ!

In fact, the performance of the main frequency of the memory is enough to affect the performance of the CPU, graphics card and other hardware, thus driving the performance improvement of the whole machine. To put it simply, this is the bucket effect.

There are shortcomings in one configuration of the computer, and other hardware cannot play the most efficient.

"So?" Lin Qi nodded and said, "Can we do this?"

"Yes!" Zhao Ruofei said, "In the past, we didn't pay attention to the frequency of the particles! Even the best semiconductor factories in the world, the memory and display capacity of the chip particles produced may be the same, but the frequency and other indicators may be different." In the past, we only paid attention to the capacity of the chip, and used it as long as the target was reached! We did not deliberately select the best to create flagship products! Now that we are aware of this, we will make a more detailed division of the chip in the future. For example, with the current 512KB memory particles, we choose the best frequency for high-end products. Other particles are used in cheap products, such as the Pangu console, and we don't need those higher frequency memory modules. ……”

"Will this affect the quality?" Lin Qi said worriedly.

"No, not really! Performance and quality are two different things. Zhao Ruofei asked, "Our screening is only based on performance, not quality." Quality is the service life and yield, which is related to quality inspection. Performance is not related to quality!"

This is related to many users, who buy very expensive machines, thinking that the quality will be better than cheap products. But in fact, this is not the case, where are many machines expensive? Some are expensive in terms of brand, and some are expensive in terms of performance. Unless the service life is specifically strengthened, the service life of the generally expensive products may not be very long.

It can even be said that from the perspective of luxury, top luxury goods are all disposable, and clothes are worn once and thrown away, without considering maintenance issues. Therefore, some people who buy the most expensive luxury clothes can not wear them after washing them once, thinking that they have bought a fake, but in fact, they may be buying the real thing.

Buy better quality products, for example, they don't fade after washing, and they can be worn repeatedly for a long time. This suggests that this luxury item may be a fake.

Because, users of top luxury goods, will not wear it a second time at all, and will not think about rinsing and long-term use. Therefore, the designers who design these top-of-the-line design products will not consider the longevity of these products.

Consumer electronics are almost the same, and they are expensive not because they have a longer service life, but because they can lead the performance experience in the next few years, and buy for a better experience in a few years.

Of course, the global semiconductor industry, especially the production capacity of memory chips, is basically concentrated in Japan, and Japan's memory chip production capacity accounts for about 75% of the global memory chip production capacity. Although it is not as crazy as South Korea's monopoly of 90% of the memory particle production capacity later, the memory particles purchased by semiconductor manufacturers around the world, at least mid-to-high-end memory chips, are basically purchased from Japan.

For example, the flash memory chips used in laptops by startup electronics companies, as well as the memory chips used in arcades and mainframe computers, are all ordered from Japan. On the one hand, Japan's production capacity is relatively large, and in addition, the price is not too expensive compared to Europe and the United States.

At the same time, this is also related to the relatively backward domestic semiconductor technology, although new start-up electronics companies continue to place orders for the domestic semiconductor industry. It has greatly promoted the rise of domestic CPU, memory, video memory, hard disk and other industries in terms of production capacity and process technology.

However, at present, a single memory particle in the domestic market can only achieve a maximum of 32KB! Such a particle is more than enough for handheld computers and even home consoles.

After all, a memory module is generally multi-grained. 32KB particles, welded with four particles, is a 128KB memory module, so the configuration in the current microcomputer industry, belongs to the mainstream configuration of memory modules.

The 32KB pellet is more chicken in the high-end market. For example, some of today's mainframe computers use 512K particles!

The arcade market has been transformed by new start-ups to be at least a decade ahead of the configuration of microcomputers. In order to experience higher-performance games, the memory chips have already started at 512KB. If you want to buy such high-end granules, at present, you want to buy it from European and American manufacturers, and you want to buy it from Japanese manufacturers. The cost performance of European and American manufacturers is much lower than that of Japanese manufacturers. In addition, it is a problem of quality, and the Japanese economy has been in recession in later generations, and various quality indicators have been falsified, so the myth of good quality and high quality has gradually collapsed.

But in the 80s, Lin Qi can seriously say that the products provided by Japanese suppliers are not only cheaper than those in Europe and the United States in terms of price, but also more reliable than those of European and American manufacturers in terms of quality.

At the request of Lin Qi, Xinchuang Electronics began to put forward new requirements to memory particle manufacturers. It is also a subdivision of the main frequency of particles, and the new venture electronics not only needs to meet the capacity standard, but also requires the reading and writing speed of the particles to be faster.

Of course, the memory particles of this era are common to the memory particle technology. Using better memory particles as raw materials can also make graphics cards with higher clock speeds.

At the same time, Xinchuang Electronics has also set up a wholly-owned subsidiary, Xinchuang Display Chip Company. The company's main business is to develop graphics card technology.

Although the status of graphics cards is not as important as that of CPUs, the importance of graphics cards must not be ignored. In the past, although the new venture electronics company attached importance to graphics card technology, Lin Qi felt that it was not enough, because, now it seems that the graphics card technology of the new venture electronics is not absolutely leading, which is also because the projects of the new venture electronics group company are too many and too miscellaneous, although it is not bad for money, but it is not focused enough.

Therefore, Lin Qi is ready to gradually subdivide the company's business. Some important departments have set up subsidiaries focusing on the main business.

Ignore the competition between different departments, and let the major subsidiaries compete with world-class competitors. The assessment of achievements is not based on the performance ranking within the group, because it is not a good orientation to make more money than anyone, which is easy to make the company fall into short-sightedness. It would be a terrible thing if everyone only focused on short-term performance and profits, and ignored long-term development.

For example, Lin Qi questioned the need to invest a large amount of money in the development of the XRM architecture CPU and Pangu host. Before the results of these projects, the performance was at the bottom, and they lost money in the long run.

Even the Pangu host lost more than a year at the beginning of the listing, and the CPU of the XRM architecture, although the performance is strong, is mainly digested within the company, and the profit is naturally not very high. Even, the profits it earns are not enough to make up for the research and development expenses of the CPU. So far, the CPU project still needs the group company to take a considerable part of the overall profit to support the continuous improvement of the CPU.

As the system becomes more and more complex, the R&D cost is soaring, and even 100 million U.S. dollars is needed to invest in R&D every year just to improve the Pangu operating system. The Pangu operating system, on the other hand, does not generate benefits and is a near-free distribution. All Pangu hosts are pre-installed for free, and even if the system is broken, buying a genuine system disk is only about $1. This price is not at all to make money, but only to charge the cost of production and make the sales channel willing to put the disc on the shelf.

These projects seem to be loss-making, but they are the core competitiveness of the new venture electronics group. The for-profit sector is equivalent to a high-rise building. And the reason why the building can be built so high is not the building that can be seen on the ground, but the invisible foundation underground!