I Found A Planet

On the 1st of May, a of great piece news came from Supercomputer Research Center in Tianjin city. Almost a year after its start, the first supercomputer independently developed by Xing Hai Technologies had been successfully produced. It was called “Supernova 1.” Now that Xing Hai Technologies had a supercomputer system that they had independently developed, they could enjoy great savings in the purchase of supercomputers.

The person in charge of the development of Supernova 1, Zhao Zhang, explained the performance of the supercomputer to Chen Jin via a video teleconference call. Zhao Zhang was a man in his 40s. He wore a long white Chinese-style gown and a pair of spectacles. He used to work as an associate professor at the University of National Defense, and had participated in the development of the supercomputer “Heavenly River 2.” He had taken a very important role in thatproject as the “secondary architect.”

Not only did Chen Jin offer him an annual salary of 10 million, but Xing Hai Technologies also had to sign a strictly confidential agreement with the University of National Defense and paid 100 million in technical licensing fees before they agreed to let Chen Jin poach him. In reality, the architecture of the Supernova 1 was brought over from the Base of Operations by Chen Jin. It was obviously more advanced than the structure of the “Heavenly River 2.” Chen Jin only poached him because he was more familiar with this aspect of supercomputers. No matter how advanced technology was, an experienced person was still needed to execute and realize it.

The University of National Defense had already developed the “Heavenly River -3E” with a computing speed of more than one quintillion Flops. The technology of the “Heavenly River 2” had obviously become outdated and unimportant. Hence, it was not a big deal if some staff from that old project were poached.

However, the 10 million annual salary offered by Xing Hai Technologies created a huge sensation in the University departments and even angered many leaders. “It’s creating a disturbance in the hearts of the armies!”

…

Zhao Zhang introduced the supercomputer to Chen Jin. “President Chen, the Supernova architecture that you’ve given us is at least two generations ahead of the ‘Heavenly River 2’ and on a similar level to the ‘Heavenly River -3E.’ Theoretically, we can develop a supercomputer with a speed of one quintillion Flops based on this architecture. As you’ve mentioned that there is a shortage of supercomputers in the company and there is an urgent need of them, we’ve completed a small-scale ‘Supernova 1.’ The peak computing speed of the Supernova 1 is around 125 quadrillion Flops. Its speed, when used in a continuous period, is 100 quadrillion Flops. Its overall standard is comparable to the ‘Light of The Mighty Lake Tai.’ Its performance is ranked as the third in the nation. Due to the advanced nature of the architecture, the performance of its power consumption ratio is extraordinary. We achieved a ratio of 9G/W, which is better than the 6.5G/W ratio of the ‘Light of The Mighty Lake Tai.’ This is comparable to that of the Heavenly River -3E. Its performance is the top in the world; it’s pushing the limits of the traditional supercomputer. With this kind of power consumption ratio, it can help save tens of millions in annual electricity bills.”

Chen Jin nodded. The most expensive aspect of the supercomputer was not the cost needed to build the hardware, but the exorbitant electricity bills! For example, the Heavenly River 2 was a large power consumption. The power consumption of the whole machine was 17,808 kilowatts. After the water cooling heat dispelling system was installed, its power consumption would reach 24 megawatts…equivalent to the power provided by the small-scale nuclear reactor. Annual electricity bills could reach 100 million.

The continuous computing performance of the Supernova 1 was three times that of the Heavenly River 2. But, the Supernova 1 still had to cut down on 30% of its energy use such that the annual electricity bill could be reduced to about 66 million. This would result in more cost-saving in its usage.

“In addition, the supercomputer architecture that you’ve given me is exceptionally stable. The failure rate of the typical supercomputer is relatively high. Whenever one computing core malfunctions, we need to spend 30 minutes to repair it. There are thousands and millions of cores in a supercomputer, hence, a couple of cores failing is a common occurrence. This type of supercomputer cannot satisfy long-term use; they can only be used for some short-term supercomputing projects. Many of the company’s program application requires long-term operations and computing. The supercomputer cannot fail on a regular basis. Our Supernova 1 can self-detect any faults, and is equipped with intelligent parallel technology. One or two malfunctioning cores will not affect the normal use of the supercomputer. It can be directly replaced when the machine is switched on. Even if half of the cores are not working, the supercomputer can still be used normally. In terms of stability, the architecture of the Supernova is not any weaker than the architecture employed by the Heavenly River 3. Moreover, the architecture of the Supernova has great potential for expansion. We can develop an E grade supercomputer based on this architecture.”

Zhao Zhang could go on all day extolling the various merits of the Supernova’s architecture. He made the right decision to jump ship to Xing Hai Technologies. Not only did the company provide an ample budget for research, but his future was also bright here too. He was very satisfied.

Chen Jin nodded repeatedly. As for supercomputers, many people group multi-cores into a simple and easy term called “heap cores” in their mind. However, the reality was far from simple. When tens of thousands and hundreds of thousands of computing cores were working in tandem to process one task, the massive data generated could easily exceed the transmission bandwidth of the communication lines as well as the read and write capabilities of the memory chips and hard disks. These bottlenecks could only be resolved through advanced architecture.

There were reasons why Z country’s supercomputers repeatedly dominated the list of the top supercomputers in the world and frequently won the title of the world’s number one. It was because their chips were not good enough. Since the 80s or perhaps even earlier than that, the researchers in Z country could only develop CPU parallel computing in a horizontal manner, using quantity to compensate for quality. This was when the country was still undeveloped. It was the only legacy that the older generation of researchers had left for their younger successors in the field of supercomputing. This was also why Z country’s researchers were still able to create the world’s top supercomputer by using their outdated domestic CPU even though the foreign countries had banned the sale of advanced CPUs to Z country.

All of the essential components used in the Supernova 1 developed by Xing Hai Technologies, such as the CPU, memory chip, and hard disks, were domestic products. In order to further reduce the selling price, Xing Hai Technologies had even made a ginormous order for supplies to build fifteen “Supernova 1″s. This order had stunned the manufacturers. Hence, the production costs of the Supernova 1 had been successfully reduced. Excluding R&D spending, and only taking into account hardware costs, the production cost of the Supernova 1 was around 500 million.

Chen Jin sighed. “Indeed, it’s better to build it yourself than to buy it.” Turning to Zhao Zhang, he said, “Assemble five sets of Supernova 1 for me. The company has need for them! After you’ve completed this task, get the team to go straight into the development of a quintillion Flops supercomputer. It will be called ‘Supernova 2’ with a computing speed of one quintillion Flops.”

Zhao Zhang nodded. “I understand, President Chen.”

…

A few days passed, and bad news arrived. The person in charge of the “Star Cloud System” project, Ma Xiaobing, said, “President Chen, the number of programmers using the ‘Star Cloud Studio’ to develop Star Cloud apps has dropped to less than 100,000 from a high of 1 million. This is a very steep drop. Google has forcefully introduced an ‘anti-firmware replacement software patch.’ Any mobile phone installed with this patch will have the bottom layer of its hardware damaged once it replaces its original firmware with the Star Cloud System. The phone will be unusable. Currently, 90% of Android phones in this country have been hit with this patch. This is an action of a gangster! They are introducing a virus! The purpose is to seal off and smother our Star Cloud System!”

This was direct proof that the rumor right. It was said that there was a backdoor in the foreign software and operating systems. It seems that this was indeed true. Not only was there a backdoor in the software, but there should also be another backdoor in the hardware’s bottom layer. Otherwise, how could a software alteration damage the hardware? As such, the national Information Security Department was currently investigating this “Backdoor Incident” and would soon issue a penalty to Google Corporation.

On the other hand, the vast numbers of programmers who had wanted to participate in the development of Star Cloud apps had their confidence in the system shaken. Many chose to give up. Since the matter had progressed to this stage, this was no longer a software issue. Regardless of how good the Star Cloud System was, there was no point to it. There would also be a possibility of system collapse as long as the backdoor in the bottom layer hardware existed. Unless Xing Hai Technologies could achieve full independence in its hardware and software production, there was no way for the Star Cloud System to succeed.

A senior programmer lamented, “The death of the Star Cloud System is a wake-up call for us. Unless we achieve full independence in terms of both hardware and software, the backdoor issue will always be a problem.”

Was the Star Cloud System really facing certain death?

Could Xing Hai Technologies only remain helpless in the face of this grim challenge?