Reborn 2005: The Internet Myth

After Wang Zheng signed several contracts, Phil International's acquisition of Facebook and PA semi was successfully completed.

Phil International actually controls 100% of Facebook's shares, and its overall business will be decentralized to WeChat.

WeChat CEO Steve is of course very excited. WeChat urgently needs content other than casual games to increase market share and user stickiness.

My boss really understands the social industry.

In the past, there was Meta for picture content sharing, and later there was Tik Tok for news sharing. It had no rivals in the social field of strangers.

Now that we have Facebook, the social attributes of WeChat have also been fully utilized.

Coupled with the online streaming media platform being developed by WeChat, it has basically filled all the gaps in social networks, and there will no longer be a threat to Midland Online and Juyou.com.

The equity of PA semi company was finally obtained at 73%. Excluding the 15% originally in the option pool and the 3% of Sutong Capital, the two co-founders currently still hold 9%.

As soon as several acquisition documents were signed, Wang Zheng didn't bother to celebrate with Peter and Jeff, and immediately drove to Mountain View with his secretary.

When a company changes hands, there will be some turmoil, and some people will be more extreme, resulting in a wave of resignations.

For Wang Zheng, the most important purpose of acquiring this company is Jim Keller. It doesn't matter to others, this person must not be let go.

Silicon-based chips have a long life ahead of them, and Jim Kaiser is the leading architect in the silicon-based chip field.

I just don’t know if Apple would have been able to develop powerful A-series processors without him.

Then Wang Zheng doesn't have to worry about it.

When he came to PA semi, Wang Zheng saw that the mood of most of the lower-level employees was relatively stable. Even though they had been told that the company had changed its boss, it seemed that the impact was not that big.

Entering the office, Wang Zheng saw that the current CEO was still Daniel Dapopul, who was also the founder of the company.

After Wang Zheng’s investigation, it was found that Dapopul, who is nearly 60 years old, is also a big player in technology. SiByte, which he previously founded, was acquired by Broadcom for US$2 billion. Jim Keller also started working at Broadcom from this transaction.

Serve as chief architect.

Dapopul did not go to work for Broadcom, but came out again and founded this PA semi. Its current main business is the development of low-power chips under the power architecture.

This direction is obviously wrong, and ARM ultimately won the battle for low-power chips.

As soon as Jim Keller left Broadcom, he returned to work under Dapop.

"Welcome to PA semi, Phil."

Dapopul had no burden about the company being controlled by Wang Zheng, and his smile was sincere and enthusiastic.

Wang Zheng took the initiative to shake hands with him and said with a smile: "We also welcome PA semi to join Phil International. I hope this acquisition will not cause you any trouble."

Dapopul smiled and shook his head. After sitting down with Wang Zheng, he said: "Phil, I have been in Silicon Valley for decades, and I am used to this kind of thing. But I don't quite understand it. Logically speaking, you should try it.

How could you acquire some Internet companies and expand your social empire to acquire a chip design company?"

Altron currently has not launched any products other than modular computer room products.

In order to cooperate with the sales of the console, the completed Siri assistant and operating system are still confidential and have not been released yet.

Wang Zheng replied: “The acquisition of PA semi is mainly to prepare for the popularization of 3G networks. Under 3G networks, more and more online pictures, streaming media and other content will be transferred to mobile phones, and our social networks are no exception.

.”

"More data transmission means more power consumption, so low-power consumption, high-performance chips will become a necessity. And PA semi is exactly what it does, so..."

Dapopur twirled his gray beard and nodded in agreement: "No wonder you succeeded so quickly in Silicon Valley, Phil, because you saw the essence of the problem."

"I firmly believe that low-power chips will become the trend in the future. The power architecture mobile chip currently studied by PA semi can already achieve a power consumption of 13W and has good performance."

"Come on, Phil, it's time for me to retire."

Seeing Dapopul getting up and walking to the desk and starting to pack his things, Wang Zheng was stunned and quickly stopped him: "Dapopul, what are you doing?"

Seeing that Popper was puzzled, Wang Zheng patiently explained: "I have no plans to change the organizational structure of PA semi, and I don't want to come over and lead the experts myself. My purpose in acquiring this company is to develop the best chips."

"Obviously, you are the most suitable CEO for PA semi, Dapopul. I hope you can stay and continue to lead this company."

"With your talent and my funds, I believe we will be able to make a super chip that changes the world. What do you think?"

Dapopul really didn't expect that the young man in front of him would actually leave the CEO position to him. This kind of thing is not common in Silicon Valley.

In most cases, after a company merger, the new boss will directly drive away the original CEO in order to gain control. If he is a shareholder, he will at least leave the management position, and then parachute in someone he trusts to take over the CEO position.

However, given the situation, Wang Zheng knows nothing about chips. In this kind of technology-based company, only technology-based entrepreneurs like Dapopul are more suitable for management work.

Of course, Wang Zheng cannot completely let go. The development of the company must still be based on his plan, and the financial power must be firmly in his hands.

Dapopul stopped packing, looked at Wang Zheng seriously, and asked: "Phil, thank you for still trusting me. I am now 56 years old, and PA semi is probably my last business. I hope

It can be as brilliant as the Blue Factory and the Red Factory. If possible, I would like to work here until I retire."

"No problem, Dapopul. As long as the company can stay on the right path, except for financial work, I will not interfere with anything else."

This kind of technical master is the talent that Wang Zheng is most lacking right now. How can he be let go?

"No problem, you can appoint your chief financial officer as soon as possible, and I will cooperate."

Dapopul also introduced the company's current business to Wang Zheng, which is mainly divided into two parts. One is self-developed chips, and the other is undertaking outsourcing design business, with annual income of about US$20 million.

However, the company's current technology accumulation is all in the MIPS architecture and power architecture, and it has basically no involvement in the ARM architecture.

Although these three architectures are all based on RISC instruction systems, different companies have different priorities in chip design.

The original version of Loongson that year was designed based on the MIPS architecture, and it was used privately without authorization... The lawsuit was not settled until 2009.

The MIPS architecture will disappear in the future. Its design companies have already switched to another instruction system called RISC-V that will appear in the future, so they will definitely not be considered.

The power architecture designed by IBM is updated too slowly and the power consumption remains high. Since it was completely abandoned by Apple this year, only its own server business is left in the market, which is definitely not within the scope of consideration.

"Dapopul, do you know the ARM company in Eagle Country?"

"Of course, this company's CPU architecture is also good, with lower power consumption, but the performance is too poor. At the same price, our customers prefer to use the power architecture, which has strong performance and not much higher power consumption."

A single-core CPU has a power consumption of 13 watts. Isn’t this high?

He had to quickly shake off this concept. Wang Zheng sorted out his thoughts and said: "Mobile phones will have more and more functions in the future, and people will use their phones more frequently. I think the ultra-low power consumption of the ARM architecture

Performance will make it more popular with users. The single-core high performance pursued by the power architecture will be eliminated by the times in at most two years."

"It doesn't matter if ARM's performance is lower. We can design more cores. If dual-core is not enough, we can use four-core. If four-core is not good, we can use eight-core. Then we can obtain a better user experience through software optimization."

This is the most common approach in the future, which is to stack the number of cores and then improve performance through software optimization.

The success of ARM is not really about how good its performance is. The most important factor is that it saves a lot of power.

It is precisely because of this that Apple has insisted on using ARM architecture CPUs. After generations of improvements, it has developed to where it is today.

Multi-core is also a future development trend. With the improvement of manufacturing processes, multi-core has become standard on mobile phone CPUs.

Of course, to build a multi-core CPU, the bottom layer of the system must be optimized, otherwise the legendary phenomenon of "one core working while eight cores watch" will occur.

And system optimization? This is not a problem for Hongmeng at all. In ten years, it will be difficult to find a single-core CPU.

Wang Zheng said very firmly: "Dapopul, the company's research and development must immediately switch to the ARM architecture. I will go to Eagle Country to obtain ARM authorization as soon as possible. This will be our best opportunity."

After hearing Wang Zheng's remarks, Dapopul seemed to have thought of something and suddenly lost his mind.

After being stunned for a while, he came to his senses and said in surprise: "The multi-core idea is completely feasible. Use the system to allocate tasks to each core to improve performance. Phil, you have come up with a great idea. I have to hurry up.

Verify it!”

Seeing Popper running away, Wang Zheng shook his head speechlessly. Anyway, the ARM architecture must be done. When the authorization is returned, it will be done even if it is not done.

As for why Dapopul was so excited, it's easy to understand.

This year's desktop CPU market is experiencing a lot of drama. Blue and red manufacturers have launched dual-core CPUs, officially starting the multi-core war.

However, neither the operating system nor the application software are following up that fast, and there is no version optimized for dual-core CPUs.

Moreover, the dual-core CPUs of this period were basically glue dual-cores, which was basically equivalent to two CPUs packaged into one.

This has led to a very strange phenomenon. When an expensive dual-core CPU runs certain programs such as 3D games, the actual effect is not as good as a cheap single-core CPU.

This situation will gradually change after Microsoft withdraws from the Vista system. By the win7 era, single-core CPUs almost no longer exist.

This is a typical example of software and hardware matching technology upgrades. Without the cooperation of the operating system, multi-core CPUs are useless except as a gimmick.

This kind of common sense about the future has a huge impact on Dapopul.

After all, whether multi-core is better or single-core is still in the verification stage, and no one knows what the direction of technology will be.

The idea of ​​​​optimizing multi-core CPU execution tasks through the operating system is probably still in its infancy.

Even Vista, which Microsoft has high hopes for, only officially added optimized support for dual-core and quad-core CPUs after two iterations.

Vista is a good system, but it was born at the wrong time.

Its requirements on hardware performance are too high, resulting in many computers unable to install it at all.

This is an excellent opportunity for Hongmeng system to capture the market.

Of course Wang Zheng will not miss such an opportunity. If nothing unexpected happens, Microsoft will officially release Vista in November this year. However, its price is relatively expensive and its initial compatibility is not very good.

At this time, if an operating system that is equally beautiful, does not require high hardware performance, and is free comes out, it will definitely attract huge attention around the world.

When the time comes, it will be paired with its own CPU, fully optimized, and dual platforms of PC and mobile phones will be launched at the same time. I don’t believe that this ecological environment will not be able to survive.

When Wang was leaving PA semi, he met Jim Keller, and Dapopul was dancing happily with him discussing technical issues. He left without disturbing them anymore.

After you have finished the CPU chip design, you have to finish the GPU chip design next.

It is different from the PC field. PCs pursue high power consumption and high performance, while mobile phones pursue low performance and high integration.

So the chip in the mobile phone field is called SoC, which means system on a chip.

It refers to a chip that contains many functions such as CPU module, GPU module, DSP module, etc. These modules can be designed by different manufacturers and eventually integrated into one silicon chip.

SoC is an inevitable trend in the development of integrated circuits and the future development of the IC industry.

It is unrealistic for a company to completely master all chip technologies, and Wang Zheng is not that ambitious now.

In terms of chip technology, he has a total of four goals: central processing unit (CPU), graphics processing unit (GPU), radio frequency chip and baseband chip.

ARM does have GPU technology licenses, but their GPUs are not very good. They only started doing it after acquiring Mali GPU Company this year. Most of this brand of GPUs are used in low-end mobile phones.

Next is the more famous Qualcomm. This rogue company’s main business is litigation, ah no, it’s Snapdragon SoC.

The Andrew GPU technology integrated within Snapdragon was developed after the future acquisition of AMD's mobile GPU department, and it is not yet available.

Of course, it's still ATI's mobile GPU division.

Now AMD and ATI are planning a large-scale acquisition, involving an amount of US$5.4 billion. Silicon Valley has been paying attention to this acquisition that is about to be completed, witnessing the rise or collapse of the red factory.

A trading king of this level has no chance to participate yet, so he can only take a look.

There is another company that makes mobile GPUs in Eagle Country, called Imagination. This company’s GPUs are well-known because of Apple. Before 2017, all the GPU architectures used by Apple were from this company.

.

It seems that this trip to Eagle Country is a must.