Chapter 232: Enter the village quietly, don't shoot
"Your company's SIC8081 is even better than Intel8086 in terms of performance. I have to say, Mr. Shirakawa, you have once again surprised me."
Michiyoshi Mazuka looked at the drawing in his hand with an interested look.
In addition to the DRAM field, is Japan's chip industry catching up with the United States in the processor field?
This is good and exciting news, at least in the eyes of Michiyoshi Mazuka.
Isn't the reason why Japan gathered national strength and invested in the memory chip industry in the past few years to reduce its dependence on the United States?
From raw materials to equipment to technology, we are catching up with and even surpassing others in all aspects.
Now that DRAM is on the verge of victory, there have been unexpected breakthroughs in the field of processor applications.
Although the processor-level chip market is not large at the moment, it has made up for some of the shortcomings to some extent.
This is indeed a happy thing for Michiyoshi Mazuka, who has a strong sense of collective honor.
"It's easier said than done. I only realized after I started that chip R&D is much more complicated than I thought."
The reason why SIC's first chip was able to complete the design drawings so quickly was that it took Baichuan Feng a lot of effort.
At Tokyo Institute of Technology, the total number of manpower involved, including professors and doctoral-level trainees, is about fifty to sixty.
These are all highly educated talents, not ordinary employees.
It is no exaggeration to say that Tokyo Institute of Technology is definitely the top science and engineering university in Japan.
With so much human and material resources invested, the front-end design of the chip was finally pushed forward.
Later, with the joining of Masayoshi Shima, who had rich product experience, SIC's chip development officially entered the fast lane.
Otherwise, Baichuan Feng optimistically estimates that if the research and development of the first chip can be completed next year, it will be considered good efficiency.
"What Shirakawa-san said is that the R&D of integrated circuits is indeed incredibly time-consuming.
It is truly rare for your company to be able to achieve this.”
Michiyoshi Mazuka felt the same way as Shirakawa Kaede.
This reminded him of those passionate years when everyone gathered together and worked towards one direction.
The sense of honor of finally achieving success is definitely not comparable to that of a product that sells well.
It has to be said that Xiao Rizi's courage to concentrate its efforts on doing great things in the chip field in the 1970s was truly impressive.
“Don’t worry, Mr. Shirakawa, if the technical drawings are reviewed and approved, Fujitsu can accept to manufacture for SIC.
But if I may be frank, the current market size of processor chips is relatively insufficient.
Especially for 16-bit processors, the price is too high due to cost reasons, and the market response may not be so optimistic. "
As an industry insider, Michiyoshi Mazuka naturally knew the general market situation of processor chips at this time.
If Shirakawa had bet on 16-bit processor chips, SIC would most likely have suffered huge losses in the end.
After all, having a good product is one thing, but whether it can be sold is another.
Moreover, chip processing is not cheap, especially the mask fee for the initial tape-out, which is surprisingly high.
In response to Mazuka Michiyoshi's kind reminder, Shirakawa Feng did not show any disappointment.
"Mr. Mazuka, this time SIC8081 is only taped out and not mass-produced." This question was answered by Shima Masatoshi.
"What?!" Mazuka Michiyoshi thought he had heard it wrong.
This...are you going to burn all the extra money for fun?
You should know that the cost of each chip tape-out is as high as millions or even tens of millions of dollars.
This conversion would cost at least 200 to 300 million yen.
For smaller companies, a single tape-out failure could even lead to bankruptcy.
Does this Mr. Shirakawa really know what he is doing? Mazuka Michiyoshi wondered if he had overestimated him before.
"Mr. Mazuka, this time SIC8081 is only used for functional verification after the design is completed.
After that, we will make a simplified design based on the existing architecture.
Then a logic-optimized version of it was released, which was between 8-bit and 16-bit.
This is the flagship model we will bring to the market this year."
In fact, Shima Masatoshi has already done the preliminary planning for this part of the work, and he is quite confident in his design.
After hearing Shima Masatoshi's explanation, Michiyoshi Mazuka was not relieved but even more surprised.
Spending more than 200 million yen just for wafer verification is truly inhumane .
However, if this is the basic logic version, it is not impossible to do so.
It is just too expensive, because each subsequent version developed needs to be taped out, and that costs a lot of money.
Michiyoshi Mazuka was amazed at SIC's wealth and at the same time surprised at its ambition.
If they dare to do so, they must have confidence in their own chips and the future market to make such a choice.
"I understand your intention. Fujitsu will fully cooperate with you."
Anyway, since they were paying, there was nothing much for Jianzhong Daoyi to say.
And with Masayoshi Shima in charge, the project should not give up halfway. Besides, he also wants to see how the SIC chip performs.
In addition to his hope to make further progress in Japan's local chip field, don't forget that Fujitsu is also one of the companies that makes consumer electronics products.
If successful, there may be possibilities for future cooperation.
"In that case, I'll leave it to Mazuka-san. Shima-san will be in charge of the specific tape-out date later."
Shirakawa Kaede came here today to pay a visit to the mountain leader. After the experience of cooperation, Shima Masayoshi will be in charge of the rest of the matters.
“Also, I would like to remind Mr. Shirakawa that after the tape-out begins, it would be best to determine the production volume of this model of chip as soon as possible.
The cost of each chip is directly related to the shipment volume, and I believe Shima-san also understands the specific difference.”
Following Mazuka Michiyoshi's reminder, Shima Masatoshi also nodded. This was exactly what he had reminded Shirakawa Kaede before.
Judging from the current progress, perhaps the early promotion can already begin.
If we ignore the design costs of chip research and development, the hardware cost after it is officially put into production is not a small number.
The hardware cost of a chip generally includes wafer cost, mask cost, testing cost and packaging cost.
Among these, the testing cost is negligible, and the packaging cost accounts for about 5%-20% of the hardware cost.
The so-called packaging is to stack the substrate, core, and heat sink together to finally form the CPU we see in daily life.
In today's era, because the technology is not as complicated as in later generations, the packaging cost is almost fixed, and the proportion is not too exaggerated.
Among the hardware costs of a chip, the real big ones are actually the first two, the chip cost and the mask cost.
There is a special proportional relationship between the two and chip production.
When chip production is small, mask cost will be the absolute majority of chip hardware costs.
For example, the mask cost is $1 million, but the final output is only 10,000 pieces.
Then the mask cost allocated to each chip head is as high as $100.
This is just the cost of the mask. If other expenses are included, the final chip cost will be a staggering figure.
When the chip output reaches hundreds of millions, the mask cost will be diluted to US$0.01, which is basically negligible.
At that time, chips will account for the bulk of hardware costs, after all, the utilization rate of wafers is not 100%.
The area of a wafer is fixed, and because of the yield rate, the number of chips that can ultimately be produced is also fixed.
Sometimes the cost of a wafer can be as high as thousands of dollars, and on average, the cost of a chip will be a few dollars.
For some special chips, the chip cost can even be as high as dozens of dollars.
Of course, this is related to the complexity of the chip processing process. The simpler the process, the higher the yield rate, and the lower the chip cost will naturally be.
Shima Masatoshi once explained the interests involved in this matter in detail to Shirakawa Kaede.
So generally speaking, Baichuan Feng understood why these two people reminded him so much about determining the chip shipment volume.
However, as for subsequent promotion, we need to wait until the chip is taped out to see its actual performance before we can make specific plans.
"Thanks for reminding me, Mr. Mazuka. I will consider this issue as soon as possible." The outsourcing matter began to take shape, and Shirakawa Feng was ready to say goodbye and leave.
“By the way, some people from Intel are visiting Fujitsu today.
They are in both the DRAM business and the microprocessor chip business.
Mr. Shirakawa, do you need them to meet? Maybe since they are in the same field, they can have good communication in the processor field. "
Just when Shirakawa Kaede and others were about to leave, Mazuka Michiyoshi revealed an unexpected news.
Intel? Baichuan Feng raised his eyebrows. They came here?
Shima Masatoshi also looked surprised, as this was his old employer.
“Thank you for your kindness, Mazuka-san, but SIC is too weak.
And the project is still in the confidentiality stage, so it is better not to disturb them for the time being."
Mazuka Michiyoshi immediately understood Shirakawa Kaede's refusal. The other party did not want Intel's people to know about the existence of SIC.
In Michiyoshi Mazuka's opinion, this is still Shirakawa Kaede's lack of confidence, or SIC does not have any products worthy of praise.
So he didn't force it, just mentioned it and then let it go.
Nonsense, Shirakawa Feng certainly doesn't want Intel to know the existence of SIC, at least at this moment.
What friendly communication among peers? Aren’t these competitors?
Although the current SIC is not worth mentioning and there are no practical products.
But if Intel pays too much attention , it would definitely not be a good thing.
Before SIC has a certain level of self-protection capabilities, Shirakawa Feng hopes that SIC will keep a low profile as much as possible.
Even if SIC's products are launched in the future, Shirakawa Kaede plans to develop slowly.
Enter the village quietly, without firing a gun.
Therefore, when it comes to meeting with Intel people, Baichuan Feng doesn't even have time to hide, let alone rush to get close to them.
As for Shima Masatoshi, even though it was his former employer, he didn't want to meet with people from Intel at the moment.
Because when he was preparing to return to China, Shirakawa Kaede had just left.
Soon after, people from Intel contacted him and hoped that he would work at Intel's Japan Research Institute.
However, after comparing the conditions offered by both parties, Shima Masatoshi decisively turned down Intel's invitation.
Just a high salary plus a little bit of equity incentives can no longer attract Shima Masayoshi's interest.
He turned down the invitation on the grounds that he was starting his own business .
Wouldn't it be too awkward if we met at this time?
So like Shima Masatoshi and Shirakawa Kaede, they both wanted to leave here immediately.
As they left, Mazuka Michiyoshi shook his head and smiled.
If Japan wants to stand at the top of the world and demonstrate its technological strength to the international community, it must flex its muscles.
In Michiyoshi Mazuka's opinion, this is a good opportunity to demonstrate their own strength.
So as soon as he finished his work here, he prepared to go and meet the people from Intel.
