Chinese entertainment for engineering students

"There is no problem with the lines below, what's going on?" A young man wearing work clothes and sweating profusely crawled out from under the workbench covered with lines, his face full of frustration.

His name is Zhang Bin, a second-year automation graduate student, and he is working on a "supercomputing" project with his instructor. This is the assembly center laboratory, where a series of achievements have been produced, and it can be said that his achievements are shining through.

Self-architected desktop computer systems and self-architected single-processor servers were all born here not long ago.

What is currently being completed is the final result of this project, a "multi-processor, multi-coprocessor workstation". This is also Wang Xu's ultimate goal, which is a supercomputer that can process audio.

The leader here is the professor of the Department of Automation. In fact, this result is a bit strange, but there is nothing to be done about it. The Department of Automation is Wang Xu’s natal family, and the Department of Automation is a versatile major, so he successfully seized the overall control of the project.

The leading mentor of the project is of course Professor Yan, Wang Xu’s mentor. This has also been his most important job in the past year or so. As for being Wang Xu’s mentor, it’s enough to just name him and give him some advice. That’s none of his business. , Wang Xu can graduate no matter what.

The final formation of "supercomputing" required a lot of effort. After discussions between Professor Yan and Wang Xu, the results were divided into several products according to stages.

From single-processor desktop computers, to multi-processor desktop computers, and finally to multi-processor servers, multi-processor multi-coprocessor workstations, acceptance appraisals are carried out in different categories.

Coupled with different versions of COS systems from the Department of Computer Science, a series of computer products were successfully launched.

Shuimu University established a Tongfang Computer Company specifically for this purpose, which was responsible for production and sales. This company should not be the same company as Tongfang in later generations, and the time does not match up.

But Wang Xu doesn't care, as long as there is production and sales, then whoever is doing it will do it.

The products produced by this company cannot be equipped with DOS systems except desktop systems, but are all equipped with COS operating system.

The computer core uses an 80386 processor, which is very different from an Intel-like motherboard. For example, the system bus directly uses the PCI32 bus, which is far more efficient than the ISA and EISA bus systems used by the standard 386.

This is a newly designed system. In addition to the imported chips that the central processor still uses, the north and south bridges, controllers and other chips are all newly designed products.

The North and South Bridge chips were a project contracted by the Department of Electronics. Wang Xu imitated the structure built by later generations of finished products, and the Department of Electronics made detailed improvements.

This is also the most important part that the Department of Electronics undertakes in this system. In order to realize this important auxiliary chip, the Department of Electronics has also joined forces with the Institute of Computing Technology of the Chinese Academy of Sciences and several Suzi Institute of Electronics.

The Department of Computer Science has even arranged several professors to conduct research on EDA software. The current level of EDA in the world is generally not high, and it is not difficult to catch up. Unfortunately, Wang Xu does not know much about this aspect, so he cannot provide an overall framework design. .

Therefore, the progress of the software slowed down the overall progress, and it was not until more than a year later that the overall usable north and south bridge chips were actually designed.

Tape-out production is completed by commissioned digitally labeled research institutes, which are two generations behind the international market and use 1 micron technology. However, there is no problem in making a north and south bridge. After all, the number of transistors is limited.

The only drawback is that it consumes more power, but in this day and age, this is not a problem at all. Problems are much more important than performance issues, not to mention just using more power.

The entire project was stumbling and was far more difficult than Wang Xu originally imagined, but it still managed to produce many original results more than a year later. This is what makes Wang Xu most happy.

He never thought that those things that were just Wang Xu's imagination could actually be successfully produced with enough financial support. At first, he thought that if it didn't work, he would simply use the North and South products of other manufacturers. That's the bridge.

The north and south bridge chips are very important on the motherboard. They are mainly used to control the communication and connection between the CPU and other devices. They are also a vital part of the motherboard. Only by mastering this chip can you truly master the design and design of the motherboard. manufacture.

Otherwise, it can only be said to be an optimization of the compatibility system.

This is also the most important reason why Wang Xu uses this chipset as a public relations target. If he can't get a central processor, he starts with the motherboard, and the motherboard requires this chipset.

Fortunately, the Mizuki Electronics Department, in conjunction with several numerically labeled research institutes, finally succeeded in overcoming this shackles and completed the development of the north and south bridge chips.

In particular, the new PCI-E enhanced 32-bit bus support has been successfully introduced in the south bridge, which is the key to externally connecting multiple high-speed co-processor cards. It also adds support for USB1.0 devices. .

This is a brand new external device connection bus standard proposed by Wang Xu and perfected by Shuimu. It was only launched a few years later. However, considering that this thing is not difficult, Wang Xu simply developed it in advance.

This thing has very detailed instructions in the skills provided by the system. Wang Xu just copied it out. What he copied was directly version 1.1, which is the 10Mbps version.

This version can already support a full range of peripherals from low-speed devices such as keyboards and mice to high-speed devices such as printers, scanners, external hard disks and optical drives, greatly simplifying the peripheral specifications of existing personal computers.

Mizuki has submitted this standard to the International Organization for Standardization and the International Society of Engineers for discussion. At the same time, it has made code-level open source in Mizuki's open source community and licensed it to hardware manufacturers around the world for free.

International standards are not something you just have to think about, nor do you just have to be advanced. It doesn’t even matter if you sell more. It depends more on the right to speak internationally. But if you sell enough, it can generally form a de facto international standard. , compatible and adapted by most hardware manufacturers.

The control chip of the USB device is also designed by the Department of Electronics. Currently, there is no mass production. It can only be an expensive laboratory product. Industrialization requires the participation of the state. In this era, the power of the private sector is not enough to carry out such a scale. industrial investment.

As a result, Shuimu University's Department of Automation, Department of Electronics, Department of Computer Science and other departments held several major scientific research achievement appraisal meetings and press conferences to announce new progress to the whole society.

It has successively released a series of computer products from chips, to chipsets, to peripherals, and finally to motherboards and complete machines, allowing Huaxia to produce truly meaningful domestic computers.

What does this kind of all-round localization mean? The above people know more clearly than Shuimu University. As long as the central processor can be produced in the next step, a domestic computer that relies entirely on autonomy may appear on the land of China. .

As a result, soon, the level of attention to supercomputing projects can be said to have skyrocketed. Many project teams have representatives from above directly stationed to observe the project progress and provide various assistance when needed.

The "supercomputing" project was initiated by the private sector, but half a year later, it can no longer be said to be a simple private project. More and more units and organizations have been involved, and the project has almost become a It is even more super-project than the "Growth Method" project.

In many specific sub-projects, the state has directly invested funds, and of course it will also occupy and enjoy some results, such as the North and South Bridges and USB. It can be said that the electronics department alone cannot handle this thing.

Apart from anything else, you can't even find a place for tape-out. After all, China has just successfully developed a 1-micron lithography machine two years ago, and it's not even ready for mass production.

But once the country takes action, it will be different. The Institute of Optoelectronics, the Institute of Computing, the Department of Electronics, and the Department of Computer Science will work together in many aspects and advance hand in hand.

In the end, the comprehensive integration and improvement of EDA software, photolithography machines, manufacturing technology, and chip development were achieved, and finally an 80% yield rate was achieved on a 6-inch wafer.

This is already a yield rate at the break-even point. As long as social demand is strong, such chip manufacturers can completely generate enough profits and enter the positive cycle of production and research.

It can be said that Wang Xu did not originally intend to get involved in chip manufacturing, but the huge appeal of the project founded by Wang Xu successfully involved all organizations related to chip manufacturing, which in turn triggered the intervention of state power.

This is the horror of the butterfly effect, you have no idea how big of a storm it will eventually be.

To be honest, in the 1990s, the cost of researching a set of photolithography machines and EDA software was not high, because the technical difficulty was not as exaggerated as that of later generations.

The problem actually lies in commercialization. If you don’t have enough commercial demand, it will be difficult for these supporting things to mature slowly.

Wang Xu dared to do this because of the rise of the Internet in the future. The demand for servers will increase explosively, and servers will not have such extensive compatibility requirements for software. This is the opportunity Wang Xu sees.

But what Wang Xu didn't see was the country's interest in the independent computer industry. This was really something that a time traveler couldn't grasp.

In the 1990s, when we kept a low profile and bided our time, a large amount of military industry demand was reduced to maintenance. There was almost no need to purchase new equipment. Things like domestic chips were directly facing competition from overseas products, and it could be said that there was no chance at all.

But the country's desire for autonomous computer systems has never been lacking. After all, the military is not the only one who has needs for confidentiality and internal use.

The problem is that there has never been any previous Chinese project that could clearly point out the direction of a new series of computers like the "supercomputing" project.

Therefore, after a series of achievements in supercomputing projects, the country’s attention can be said to be increasing layer by layer. The more achievements, the greater the investment, and the greater the investment, the more achievements, entering a cycle of positive incentives.

A little more than a year and a half, which is neither long nor short. With Wang Xu’s detailed planning and detailed design instructions, although the supercomputing project has setbacks, compared with ordinary research, It can be said to be smooth sailing.

The preparations for various subsystems have been completed, and now comes the final assembly and testing process.

Several graduate students have been busy in the laboratory for several days. All kinds of messy problems can be said to be endless. Maybe it is destined to be hard work. The first few milestones went smoothly, but when it comes to the final product, everything fails.

The professor and supervisor looked at the finished product and thought carefully. Several graduate students checked the various connection interfaces over and over again, and even took out a multimeter to measure the voltage of each pin one by one.

After eliminating hundreds of problems such as insufficient power supply, weak bus interface welding, board interface manufacturing errors, etc., finally, on a sunny summer day, the final product of the supercomputing project, a multi-processor, The coprocessor workstation really lights up.

"Beep, beep, beep." This is the self-test sound of the motherboard's built-in small speaker. Three consecutive beeps mean the self-test has passed.

"Beep, beep, beep." This is the detection pass sound of the co-processing board.

"Bright, bright, bright." Well, this is the cry of the expectant graduate students.

"It's on. Can the keyboard be used?" Graduate student B asked.

"It works, let me try the basic application." Graduate student A got up from the ground, sat down on the computer chair, and looked at the screen.

"Editor, no problem, write a Hello world." Graduate student B is coming up with ideas. There is only one keyboard, but anyone can see the screen.

"It works, the test passed."

"Run the test program and test the coprocessor boards separately."

"It's running, the first version of the card is normal"

"The second board is normal"

"third……"

"Stop, don't deal with these formalities, just shout at the end, we are not Xichang."

"I didn't say anything, so you had to press me to ask. I just told you kindly, but you actually called me a formalist."

"Your words are too monotonous, first, second, third."

"If you don't call it that, why don't you call it A, B, C, D?"

"We can consider setting a standard in the future so that it will be clear for testing."

“Get it done first, then talk about it, and then improve the details later.”

"All four co-processing cards are running normally."

"Run a test program and see."

"Let me look for it. Does anyone have the program to test the coprocessor operation?"

“It’s not available on Git”

"This machine is not connected to the Internet yet"

"Oh, I forgot."

"I have it here. Yours is a 1.44 floppy drive? Hell, my disk is a 5-inch one."

"That machine over there has dual floppy drives. Go ahead and flip it around."

"hurry up……"

"Here we go, done."

Well, this is the daily routine of the laboratory. It is like this whenever the results are about to come out. Don’t talk about rules, quietness, or prior arrangements. Anyway, no research can really be done step by step. It seems that research is messy. normal.

The small laboratory can be said to be full of people and chaos. There is no sense of ceremony for the successful development of new products. It feels like a busy experimental site.

"Done."

"Successful."

"More than six thousand points, why does it feel a bit low? It can't reach four times. This is four cards."

"Four times? Who is such an amateur? You are new here. If you can get twice as much, you are a success."

News of the success of the new machine quickly spread throughout the department, and countless teachers and classmates who received the news came to share the joy of this success up close.

This news had to be confirmed by several supervisors before it could be reported to the superiors, but internally, everyone got the news within almost an hour, including Wang Xu of course.

——

(See you tomorrow, finally caught up)