16
It took Wang Xu more than a week to roughly categorize all the technologies required for a digital camera. That's right, just the categorization took him over a week.
This is the complexity of modern equipment. Not to mention anything else, the casing alone uses many different materials, from plastic to rubber, from metal to non-metal—it’s a complete assortment.
Wang Xu could understand that the material for the handheld grip and the sealing parts would definitely be different, and that some internal brackets needed various supports, so using metal was fine.
But what was the logic behind using aerospace-grade titanium alloy for a camera's battery bracket? It was just a thin plate, providing basic support and perhaps a bit of heat dissipation, nothing more.
In Wang Xu's view, copper or aluminum could have solved the problem, but they still used titanium alloy. If Wang Xu hadn't carefully scrutinized it, he would never have discovered such a big surprise hidden there.
But the surprise, of course, didn't come from the titanium alloy itself, but from its processing technology, because this item was 3D printed using laser additive manufacturing.
The precision was quite high, and it was formed in one go. The working principle and basic structural diagram of the processing equipment also existed, but there was no further detail.
"If only this thing could be infinitely recursive, then we wouldn't need to buy any technology in the future," Wang Xu thought, a bit greedily.
However, this thing was indeed a good find. 3D printing was actually invented as early as 1986, and even commercial products had emerged.
But currently, these devices could only print materials like plastic or resin. Although metal printing had theoretical backing, no actual equipment had been developed yet.
Moreover, patent issues were quite annoying because the original inventor applied for a very foundational patent that was almost impossible to circumvent.
In later generations, it wasn't until this patent expired in 2005 that 3D printing truly began to develop on a large scale worldwide.
However, for Wang Xu, it wasn't a big problem. Wang Xu estimated that he wouldn't sell 3D printers; if he developed some equipment for his own use, no one could control him.
Seeing a good technology, he would extract it, write a report, explain the technical parameters and indicators, and then bind it into a book.
Wang Xu spent those days doing just that. His two secretaries were kept busy. He would dictate, and the secretaries would type it into the computer.
Some things also required drawings. Wang Xu would make a simple sketch, and then the secretaries would find someone to draw it using software like CAD.
As for chips, only basic descriptions could be made. The internal design diagrams were too complex. Unless Wang Xu meticulously drew them piece by piece with software, he could only choose to find someone to develop them.
Moreover, the process required for this thing already exceeded the current limits of global manufacturing processes. Even foreign countries didn't have such equipment, so equipment also needed to be developed.
These things could only be included in the five-year plan. But that was fine; this was originally a major item in the plan.
Especially after Wang Xu created a series of software and hardware products that caused a huge stir worldwide, investment in this area significantly increased.
Starting in the mid-nineties, the country began to accumulate some wealth, and thus gradually started investing in a number of major projects.
It was like this in his previous life. After 2000, these projects began to yield results, and around 2010, there was a concentrated explosion of output.
Many of the black technologies we see in later generations began to receive investment during this era.
The earliest, of course, were various infrastructures, such as large-scale petrochemicals, oil, steel, non-ferrous metal smelting, and railways.
And there has always been a significant amount of investment in the research and development of computer software and hardware. In fact, there have been achievements, but they seem somewhat insignificant compared to the massive international investments.
What's more, the Hanxin incident caused a lot of trouble, making it so that for many years, no one from top to bottom dared to touch this area.
However, everything is still good now, and there are still many ambitious plans, especially Wang Xu's series of major moves, which directly stimulated the entire industry.
Jiutian—oh, the company founded by the Digital Research Institute that initially developed the Northbridge and Southbridge—has already started tape-out for its RISC co-processor, a general-purpose floating-point computing chip using the open-source RISC-1 architecture from GIT.
This is not a CPU, but closer to a later-generation GPU, just without some preset graphics and image algorithms. It's a very streamlined floating-point arithmetic unit, designed specifically for floating-point operations.
It is also an outpost for their ambitious CPU plan. The purpose of designing this was because ever since the 486DX, Intel integrated the co-processor into the CPU, specifically for handling floating-point operations, and it used a complex instruction set.
This was very unfavorable for the continued development of Wang Xu's stacked architecture. That's why he discussed this with Wang Xu. In the future, whether it's for stacking Supercomputers or graphics workstations, it can be adapted through software to push floating-point operations to the stacked auxiliary boards.
With this method, by simply adapting OpenGL, it can be fully utilized as a 3D graphics card.
It's just that the RISC model requires more software development, but the software is available on GIT, and it was originally adapted for RISC.
Strictly speaking, it's not impossible to use this directly as a CPU, but it would require re-adaptation, and due to the lack of an integer processing unit, it would need simulated operations, making it less cost-effective.
For Wang Xu, this product was like a timely rain.
The new generation Supercomputer project will use Intel's Pentium processor or IBM's new generation PowerPC chip as the main chip, i.e., the CPU.
All required floating-point operations will be pushed to the array of new chips for computation.
This structure makes the entire computer's demands on the CPU very low; even using some older chips is fine, as it only serves to provide basic operational capabilities.
More complex computations won't use it, which, of course, requires the operating system to support this from the bottom layer.
The latest version of the COS kernel, as well as various Science computing libraries on GIT, already provide such support.
What does this indicate? Of course, that the second-generation workstations are already on their way.
Don't let the fact that the first generation came out only a year ago fool you; Wang Xu believes that by early next year at the latest, the second-generation workstation and server systems will be complete and become commercial products.
This time, it is no longer a project solely led by Wang Xu. Although Wang Xu himself still plays an important role in the project, countless researchers have truly grown.
Inspur, Founder Group, Tsinghua Tongfang, Great Wall, Jiutian, and even Huawei and ZTE, these companies have all played important roles in this.
Don't be surprised; telecommunications companies have also joined in. Communication between stacked computing units and the host is inherently a major challenge in the project.
The biggest difference from the development of the first-generation Supercomputer project is that this time, commercial companies are almost solely responsible for the main effort.
The computing institutes and research units that originally undertook important R&D work have almost all spun off their R&D in this area and are operating as independent companies.
This is the biggest advancement of this Supercomputer project. Researchers can now earn higher compensation from commercial companies than before, without having to go through the complex approval processes of the past.
Plans can be decided within commercial companies, without needing various project approvals or other bureaucratic hurdles.
Don't underestimate the server and workstation market; its international market size is in the hundreds of billions, a scale large enough to sustain a multitude of companies mining for gold within it.
Wang Xu was very pleased to see such changes. He even felt that for the next generation of development, his intervention might be completely unnecessary; he would only need to replace machines when the time came.
At least on the hardware side, this is certainly the case. On the software side, Wang Xu still maintains control over a few core projects, which require updates periodically.
But that's all. He basically no longer needs to re-adapt for the new things these companies come up with; he only needs to check their code.
Clearly, this cornfield is no longer being harvested by Wang Xu alone. There are no fewer than twenty to thirty universities involved, and hundreds of companies.
This is just domestically. In fact, due to the openness of this system, although there hasn't been international cooperation this time, it's almost certain that foreign enterprises will be invited to participate next time.
The increased support for PowerPC this time is precisely because old guard IT industry giants like IBM are willing to step in and support it.
Sun Microsystems and SGI have been somewhat agitated recently. As the software on GIT matures, their proprietary ecosystems, which once served as barriers, are showing signs of gradual collapse.
And GIT's support for workstations and servers clearly prioritizes COS, which has led to a large number of potential customers beginning to switch to the GIT ecosystem.
If it were just Chinese companies, it would be fine, as their products have almost no presence in the United States. Their current sales in the United States are still less than a fraction of their total.
But some traditional PC manufacturers in the United States, who had never ventured into this field, are also sharpening their knives.
Dell, Compaq, and even the struggling AST, which is on the verge of collapse, have started launching workstation and server products based on the same architecture.
This cannot but make these traditional server and workstation manufacturers anxious.
Under the impact of this system, the barriers between minicomputers and personal computers have almost disappeared. Operating like microcomputers, computing like minicomputers, and even medium and large-scale machine products are emerging like mushrooms after rain.
What's even more frightening is that Europe and Japan also seem to have seen an opportunity.
Throughout the 1980s, Europe and Japan suffered an almost complete defeat in the personal computer era. Not only were they unable to export, but even their own markets were disrupted by the personal computer trend.
In the face of the powerful WinTel alliance, the entire Europe and Japan suffered a complete defeat, from hardware to software.
It seemed the whole world used American computers, American operating systems, and American application software.
The emergence of GIT and the budding of a new architectural system drove a wedge into this strong combination. It's worth noting that COS also started with Intel chips.
Intel had almost no reason to attack an operating system that focused on various Science computations.
But recently, the tide seems to be turning. Some individuals using personal computers also seem to be shifting towards various graphical, more convenient releases of COS.
Initially, due to COS's higher hardware requirements and the necessity of expensive co-processor cards, or even card cabinets, for smooth operation, its equipment was too bulky and unsuitable for home use.
Only a small number of enterprises and departments with scientific computing requirements used COS for office work; the vast majority of users remained with the WinTel alliance.
But since the latest version of the COS kernel was released in December, this version added support for the latest Pentium processor, and could even utilize its built-in floating-point co-processor to handle some basic graphics and multimedia operations.
It even provided software support for the complex multimedia MMX instruction set, which seemed to suddenly open Pandora's Box.
Because this meant that graphical operating systems, smooth and simple mouse operations, various familiar elements of later generations like the taskbar, notification bar, Start button, right-click menu, single-click, double-click, scrolling, and so on, all appeared in the initialized COS graphical interface.
The initial graphical interface of COS was not rudimentary, but it always had many problems. It had been a year, and these bugs hadn't been fixed.
This was mainly due to the lack of commercial company operations, but now, if large hardware manufacturers start supporting such an operating system, it means its flaws will be quickly resolved in a very short time.
No one would doubt the power of these large computer companies. And Microsoft's new operating system is still in testing, far from launch, so during this window, Microsoft faces an unprecedented strong challenge.
Wang Xu knew all this because Director Jiang told him. Since the last meeting, the work of transferring GIT to various universities has significantly accelerated, and relevant departments have also clearly increased their attention to the various data on GIT.
They were surprised to find that the data on it could, to some extent, analyze many overseas business activities.
For example, recently, they have clearly noticed that the update frequency, number of contributors, and nationalities for certain GIT projects are all significantly increasing, which allows them to analyze the trends in the Western IT industry.
Suddenly, this thing was no longer just a technical open-source platform; it had almost become a compass for global IT trends. Sociologists and economists were already conducting specialized research on it.
It is said that the Wharton School in the United States is collaborating with that secretarial school in Shanghai to study the relationship between various dynamics on GIT and the IT economy. They seem interested in establishing a set of indices to indicate the direction of IT development.
This is the power of GIT's internationalization. All information is public, anyone can research it, and international cooperation and competition show surprising changes on a small GIT.
This was definitely not something Wang Xu could have imagined when he open-sourced it. Perhaps the hurricane caused by Wang Xu's butterfly wings in this world is manifested through GIT.
—
(Starting work now, see you tomorrow)
