Chapter 219 - 213 Difficulties_1

The shuttle program under the FATS plan did not restart smoothly, and the same was true for the SLS.

After a month of overtime efforts, the new core stage of the SLS rocket was designed and delivered for manufacturing, but the SLS rocket still had numerous problems.

This time, its enemy was cost.

As the successor to the Shuttle program, the SLS rocket aimed to be cheaper. The initial cost per launch of the Shuttle was about 500 million US dollars. The goal of the SLS plan was to have a greater payload capacity than the Shuttle—expanding from over 30 tons to 130 tons—with a cost of about 2 billion US dollars per launch, a higher cost-performance ratio, and lunar landing capability.

The dream was grand, but reality was skeletal.

The price of each of the four RS25 engines used in the SLS core stage was 75 million US dollars. Combined with the extremely expensive five-stage SRBs (solid rocket boosters), Director Claire recently estimated the cost of launching the SLS for the first time by the end of the year. Including the expenses for research and improvements, it would definitely not be just 2 billion US dollars.

At least double that amount, more likely more than double, around 5.5 billion US dollars.

5.5 billion US dollars just to launch a rocket?

In other words, the early stages of the Artemis plan’s three SLS rocket launches would cost 16.5 billion US dollars?

The price of two rockets exceeded the entire budget for China’s Dawn lunar program!

Upon discovering this fact and the budget for renovating the Endeavour, Director Claire quickly covered up the news, definitely not allowing it to leak out.

The only option was to pretend to be ignorant, waiting until the project was too far along to stop, forcing Congress to continue investing and keep the whole project going.

NACA had plenty of experience with this kind of situation, no need to worry.

...

"EDA software design is also provided by New Yuan."

On the fifth day of the semiconductor summit, July 26, Wang Minjiang once again uttered the sentence that made everyone’s scalp tingle.

The main purpose of the semiconductor summit in 2016 was single: to transition to ternary technology, surpassing the traditional.

The first and second days of the summit were mainly for Yellow River Semiconductor to introduce how far New Yuan had progressed with ternary technology.

Aside from a few semiconductor manufacturers who had begun contract manufacturing and thus had some suspicions, the rest were shocked that New Yuan had quietly advanced ternary technology to such an extent.

For chip design, there was one. New Yuan’s B-level Base’s automated silicon carbide production line was equipped with specialized EDA design software, very mature and stable.

In order to adapt to traditional silicon wafers and analog FETs, New Yuan’s Microelectronics department made significant improvements to this particular EDA software at that time, which was later used by SinoCore International to produce the masks for lithography.

For the foundational environment, there was one. From the SC09, the Microelectronic department deduced a set of high-level languages belonging to the Ternary System, internally named X language.

X language was different from the popular programming languages at the time; it was a mix of Chinese and English.

At the fundamental logic level, X language supported both Chinese and English. Most instructions were resolved using a single character, while some complex instructions could be replaced with letter abbreviations; it did not require the programmer to have a basic level of English proficiency. As long as they could remember some code abbreviations, they were good to go.

If they were used to English code, they could still use it. Since it was based on ternary FETs, for ternary computers, there wasn’t an issue where a Chinese character occupied more space than an English letter.

In ternary computers, both a character and a letter took up 1 byte.

At that time, Wang Minjiang also brought two laptops using the X32115 chip. These were designed for engineers on business trips with satellite internet notebooks, looking much bulkier than mainstream laptops; however, their primary function was only for internet access and document editing, where performance was fully sufficient.

In fact, the X32115 was powerful enough to compete with the mainstream CPUs of the time, and even had graphic computing capabilities—it’s just that there were too few software options, and the system was rather average, barely put together well enough for use.

These two laptops, from the CPUs, memory, storage chips, and major components, were all domestically produced, and loaded with New Yuan’s internal ternary programming software, brought for demonstration purposes.

During the demonstration, Wang Minjiang briefly explained X language, then invited a few of the attendees to try compiling programs on the spot.

This bilingual programming approach felt quite peculiar to them. Although the code written during the experience was no more than small entry-level tools for learning, they could still perceive the maturity of X language.

The high efficiency wasn’t immediately apparent, but it was clear that X language programming wasn’t verbose. At the very least, it was comparable to the mainstream C family languages and Python, and quite easy to get started with and learn quickly.

In contrast, Easy Language, due to its founder’s closure, had fallen behind the times many years ago, not even supporting X64 architecture—nobody used that antiquated thing anymore.

The operating system that New Yuan stripped down and simplified from the SC09 was quite average in performance, but it ran smoothly without any bugs. It just lacked a software ecosystem which was also quite a pleasant surprise.

Over the five days, the confidence of the businesses attending the summit grew from none to increasingly inflated, even beginning to swell with ambition.

Originally, they thought ternary computing was very unreliable, but New Yuan had already mastered chip design, fundamental logic, operating systems, program compilation—a basic whole set of processes—albeit somewhat monotonous.

Although they didn’t reveal any information about their most attractive XW series system equipped with AI, the potential within was sufficient for people to understand.

For any difficulties encountered, Yellow River Semiconductor, which was authorized with a plethora of patents by New Yuan, only had one sentence to say, "We have it."

This wasn’t about technological breakthroughs, it was a platform set up and waiting for them to perform.

Over the course of a few days at the summit, Modu Huahong announced that it would focus on the manufacturing of ternary memory and storage chips; Unigroup and Loongson decided to develop ternary architecture desktop and mobile CPUs; Jiangcheng Digital Engineering Research Institute planned to immediately research ternary architecture graphics processors and AI processors.

SinoCore International... SinoCore International looked towards Modu Microelectronics—or rather, everyone present looked towards Modu Microelectronics.

He Yuming, the general manager of Modu Microelectronics who attended the summit, felt somewhat embarrassed because if this summit served as an encouraging tonic for IC manufacturing and design companies, it was like a death knell for Modu Microelectronics.

After five days of rigorous discussion, everyone came to a unanimous conclusion: even with the full exploitation of the advantages of the Ternary System, it was only possible to circumvent the difficulty of lagging processes on a small scale.

Now 28nm ternary roughly equated to 14nm binary, and if the world’s mainstream entered the sub-10nm era, the Ternary System would at least need to reach 20nm or 18nm.

That is to say, it’s 2016 now, and by 2020, domestically produced lithography machines must achieve a minimum processing node of sub-20nm.

If not, then the shift towards ternary would be considered a failure, still unable to reverse the backward situation.

The current 90nm lithography machines weren’t up to the task, nor would the upcoming 65nm ones be enough; what was needed were 40nm or 40nm++ lithography machines to barely reach 14nm processes.

But could He Yuming guarantee the delivery of a 40nm lithography machine before 2020?

It wasn’t just the lithography machines—when it came to processes inside of 20nm, the requirements for chemical materials also rose significantly. Could the domestic industry keep up?