Chapter 770 OEM chips for AMD
In the 1990s, after Moore retired, Grove took over his position. He originally said that he would retire at the age of 57, but he worked very happily and was reluctant to do so. Until the past few years, Intel continued to decline, and he could only Able to retire.
Berrett took over his duties, but what he faced was a mess.
Under the halo of Loongson, Intel, the processor giant, is only left to survive.
They originally thought their competitor was AMD, but they didn’t expect that the little-known Dongfang came out with a dark horse!
In fact, Intel was also engaged in memory a few years ago, but it suffered losses due to fierce competition in the early 1990s and withdrew from the market. Now, although they do not produce memory, they can specify standards.
“Yes, their use of DDR is the wrong route. This technology cannot bring about substantial improvements. We have already negotiated with MRambu Company to use their standards.” Technical Officer Paul said.
He is confident: "They have evolved from the PC100 route. The memory bandwidth has increased from 1GB/sec to 2GB/sec at most, which is the PC133 standard at most. The data bandwidth improvement is limited, and our RDR memory standard has an initial bandwidth of It can start from 4.2GB/sec, and the development prospects are unlimited. It should be a perfect match with our new generation processor. Even if our main frequency is not as high as theirs, our computing power will not be slower than theirs. How many."
Paul is responsible for memory specifications. He went to Rambus many times and jointly customized the Rambus DRAM memory specifications. After all, everyone saw that existing memory was about to experience bottlenecks and required a new memory standard. This kind of memory uses RISC architecture internally and has high performance. The advantages of frequency and high bandwidth can definitely become the new darling of the memory market. I thought that Loongson, which is good at innovation, would come up with any new standards, but it turned out that it was just modifying the original memory!
In terms of frequency, Rambus memory also has great advantages. It directly uses a 16-bit bus of 400 MHz and can also transfer data on the rising and falling edges. The starting frequency is 2400 MHz.
(If compared, it is similar to the memory sticks in the DDR4 era, specifically DDR4-2400. This shows how advanced the technology of this memory stick is. Unfortunately, advanced is useless. The key is the price. Who can afford it if it is too expensive.)
“Then what are you waiting for?” Barrett said: “Contact all memory manufacturers now and ask them to unify their stance and refuse to support DDR memory.”
Paul shook his head helplessly: "Mr. President, you have just taken over, so you probably don't know yet. We have reached an agreement with Rambus. This kind of memory technology is only owned by us and Rambus. We have no intention of licensing it to third parties. "
They publish new standards, and for what purpose? They still want to squeeze out all other memory manufacturers and let them dominate the market. In the past few years, they continued to lose money in the fierce memory competition, and were finally forced to withdraw from the memory field. Now, they want to bring new standards. Returning to the stage, no one wants to produce, Intel and Rambus are left alone!
"Now, can there be a future without authorization?" Barrett glanced at him: "First find ways to squeeze out Loongson Company. With them here, none of us can expect a good ending!"
If you want to live alone now, you will definitely not end well. You will definitely go bankrupt. In order to develop, they can only join forces!
Faced with this challenge, Intel's choice is to continue to compete. After all, they were once the leader in this field, but this is not necessarily the case for other manufacturers.
Such as AMD.
Unlike Intel, AMD has not changed its leaders since it was founded in 1969. This is a miracle among high-tech companies, and Jerry Sanders has a legendary spirit.
Generally those who engage in technology come from wealthy families, but Sanders is different. He was born in 1936 on South Winchester Street in Chicago. When he was a child, his parents divorced and he could only live with his grandparents.
Long-term poverty life has also created Sanders' perseverance. Once he decides on something, he will move on unswervingly.
When AMD was founded, there were only eight people. Sanders wore many hats, especially to defraud investments.
Intel, a promising company, took five minutes to raise $5 million, but Sanders took five million minutes to raise $50,000. However, he did not give up because of this, he kept fighting.
Relying on an unyielding spirit, AMD survived the difficult early stages of entrepreneurship. However, their future is still unclear. After all, they are facing a behemoth like Intel. Sanders has never been afraid. He took AMD to challenge Intel. He also started from himself, improving the company structure, stimulating the innovative spirit of employees, etc. It can be said that it is precisely because of such a resolute and decisive person. Only outstanding leaders who dare to speak out and never give up can maintain a high morale at AMD. Look at the companies that challenged Intel together. Zilog, National Semiconductor, and Motorola have all disappeared in the long river of history.
Saunders is now gray-haired, in his sixties, and has reached retirement age. However, he has to lead AMD out of trouble. There are not many paths he can choose.
"Now, our biggest problem is that the manufacturing process cannot keep up." Dirk Mayer said: "Like Intel, we have just mastered the 0.18 micron manufacturing process and cannot maturely use it. Our K7 also plans to trial it first. 0.25 micron to make the transition. SMIC has long been able to use the 0.13 micron process. By the time we master the 0.13 micron process, they will definitely have mastered the 90 nanometer process. They will be able to beat the 90nm process by relying on technological progress. us."
Mayer only joined AMD in 1996, but he is definitely a powerful person. He was hired by AMD from IBM. He served as engineering director when he arrived and was responsible for the AMD-K7TM microprocessor development plan in Austin, Texas. . Originally, this project should be very successful. The initial model of the K7 processor can reach a frequency of 900 MHz, and will eventually reach 1.8GHZ. Its appearance allowed AMD to defeat Intel for the first time, making people shout AMDYes slogan comes.
Mayer was promoted to vice president of engineering for the Computing Products Group on the basis of this project, and it was in this capacity that he came to the meeting.
But, so what if this processor beats Intel? A more formidable opponent than Intel has appeared, and they are suppressing Intel!
"Yes, the more advanced the process, the more transistors can be integrated on the same chip, the lower the cost, the higher the performance, and the smaller the heat generation." Sanders said: "We cannot rely on 0.18 microns. To win through technology, we must continue to develop 0.13 micron technology."
"On our own, we won't be able to master this production process in less than three to five years, and by that time, our opponents will have already switched to 90 nanometers." Mayer said: "We will always follow, and we will always follow. They all have no future.”
AMD also has its own wafer fab. They design and manufacture by themselves. If there are problems, they have to solve them by themselves. Therefore, their progress is not smooth and their steps are difficult.
“What should we do?” Sanders asked.
"Our opponent, Loongson, is only responsible for designing processors. They can spend all their energy on design. As for how to produce processors, it is left to SMIC. The latter does not design chips, but only produces chips. ." Mayer said.
“Whatever you want to say, just say it directly.” Sanders said with a dark face.
“We improved the design of K7, adopted a 0.13 micron process, and then handed it over to SMIC for foundry. In this way, we can directly keep up with the pace of the East.” Mayer said.
“This is impossible!” Sanders categorically rejected: “Heroes need to have their own fabs” (realmenhavefabs).
This sentence has always been circulated in Silicon Valley. It belongs to Sanders. He is a famous tough guy in Silicon Valley. Like the founder of Intel, he came from the Whampoa Military Academy in Silicon Valley: Fairchild Semiconductor Company, but he fought with Intel. for more than thirty years.
For him, if it doesn’t have its own wafer fab, is it still a high-tech company?
When Sanders was in office, he always insisted on independent research and development and independent production until he retired. His successor Lu Yizhi sold AMD's wafer fab business in 2009 and established the Middle East Rich Fund. The current Globalfoundries company (that is, GF, which is very similar to the abbreviation of Girlfriend.)
However, despite being split up, AMD still handed over its chips to its girlfriend for OEM production, and after receiving investment from wealthy people in the Middle East, they gained a financial advantage and made rapid progress.
It wasn’t until 2011 that my girlfriend’s 32nm process foundry problems caused AMD to suffer, so they handed over some orders to TSMC for foundry. By 2012, AMD completely broke up with his girlfriend and completely gave up his negligible shares. Since then, AMD has completely embraced TSMC, relying on TSMC's process advantages to once again beat Intel to the ground. So much so that computer enthusiasts have to thank AMD for allowing them to use cheap Intel.
Now, AMD has encountered a crisis, and their only solution is to OEM!
"We are not going to completely abandon our wafer fabs." Mayer said: "Our wafer fabs also need to develop. However, our technology cannot keep up with the Easterners. If we want to compete, we have to use the same process. Competition, otherwise, the problem is not whether we have our own wafer fab, but whether we have AMD itself."
If this continues, AMD will also go bankrupt!
"Mayer, you are taking it for granted. How can Easterners give us OEM work? They are eager to squeeze us out." said an executive.
"If the people in the East are smart enough, they should see that there are many benefits to OEMing for us." Mayer said: "If their processors want to be sold here, they must abide by our laws. In the world, There can't be only one processor company or it would violate our monopoly laws."
Everyone is talking, and Sanders is thinking about it. From his personal point of view, he absolutely does not want to hand over the chip to others for OEM. However, now that their own technology cannot keep up, it is no good if they do not hand it over to Easterners.
The wrong route is holding them back. If they switch to immersion lithography technology early, how can they let the Easterners take the lead? Are all the companies that make photolithography machines just suckers?
"Our goal is not to compete with Loongson. At least for now, it is difficult for us to surpass Loongson. However, we only need to defeat Intel." Mayer was still talking: "Our national defense, government, Education, etc., many government departments need to update computers, and they will definitely not buy Loongson computers, even in Europe. If we defeat Intel, our AMD company will dominate this market. When the time comes, we will continue to grow and develop. Ten times, a hundred times. Let’s develop first, and when we close the gap with the East in the future, it won’t be too late to overtake us.”
These words finally made Sanders make up his mind: "Okay, then we will go to the East and apply for OEM work from them."
Chip foundries must receive a complete set of design drawings, so their processors have no secrets!
If you use the same foundry as your competitors, will the foundry leak their secrets to Loongson?
“Our technology is not as advanced as Loongson’s, so why would they care about our secrets?” Mayer said, “Moreover, we use different technical routes and different instruction sets, so what can we reveal?”
“You’re talking about now, what about the future?”
"In the future, they will find that there are huge profits from contract manufacturing for us. We will take advantage of this opportunity and invite them to build factories in the United States, so that we can firmly control them." Mayer said: "Even if we want to obtain White House authorization, those smart alecks, shouldn't stop us on this matter. They will focus on Intel and let us fend for ourselves."
Saunders no longer hesitated: "Mayer, you will go to the East immediately to discuss OEM matters with them."
“Me? Aren’t you going to go in person?”
“I’m going to the White House to lobby.”
In this way, Mayer embarked on a trip to the East. In fact, he had no idea. If the opponent was a person who did not have an overall view, he would definitely not do the work for him. However, if the opponent was Mr. Qin, If so, he should be able to know how to make the right choice.
Time has entered 1999. At the beginning of the new year, everything is new. Many things have to be arranged, especially the China Iron and Steel Association to discuss the iron ore imports in the new year. Qin Tao is even more busy. He has been in the capital and China Iron and Steel. The leaders of the association were mixed together. On this day, Zhao Ling walked gently to his side while feeling dizzy.
“Brother Tao, there’s a new situation at SMIC. Mr. Zhang can’t make up his mind, so I’m here to ask for your opinion.”
"what's the situation?"
“AMD came to us to produce K7 processors.”
"What? Looking for us to do OEM? What a good thing!" Qin Tao said: "Of course we have to agree to such a good thing! Tell Mr. Zhang that without affecting our own production, the excess production capacity can be given to AMD. work!"
After saying that, Qin Tao thought of something else: "I'll go there myself."
“Brother Tao, they are not the CEO, they are just the people under the CEO.”
“It doesn’t matter, we just want to make the other party feel that we value them.”
Zhao Ling didn't understand even more: "Brother Tao, isn't it good that we crushed all our opponents? Why do you pay so much attention to them?"
"Of course it's good to crush our competitors, but what happens after we are crushed?" Qin Tao said: "Will others let us go? The other side will use all kinds of dirty tricks. If we can OEM for AMD, it will be very good." Yu Duo gave SMIC a way to survive. However, you must be careful in everything to avoid being plotted."
Qin Tao went to Huating from the capital again. When he arrived at Huating, it was already evening.
The talks could only be held on the next day. That evening, Qin Tao met with Zhang Rujin first.
“We are now producing chips, do we have the entire industrial chain in our own hands?” Qin Tao got straight to the point.
"Producing chips requires wafers. The equipment includes single crystal furnaces, vapor phase epitaxy furnaces, oxidation furnaces, and magnetron sputtering stations. To make wafers into chips requires chemical polishing machines, ion implanters, photolithography machines, etching machines, and scribing machines. Chip machines, etc., we can currently produce these equipment ourselves, but the output is not high, single digits every year, so it limits the production scale of our own factory." Zhang Rujin said: "When purchasing equipment, we follow the principles If you can use domestically produced products, you don’t need to import them, so in the past few years, domestically produced machines have become available, forming this complete industrial chain.”
“If you think about it again, is there anything else?”
“Others, it’s photoresist, which we still import from island countries,” Zhang Rujin said.
Although Qin Tao is not engaged in electronics, he is deeply impressed by this photoresist. Such a small thing may be stuck by others. Relying on imports is unreliable. The little devil has to listen to his father. Moreover, this thing is different from iron ore. If you import too much at one time, if it is not used, the performance will be reduced after six months and it will be broken after nine months. Yes, it has a shelf life.
“Can’t we produce it domestically?”
"Low-end photoresists can already be produced by ourselves. Relevant domestic units are developing high-end photoresists. We are also constantly testing, and we are getting closer to success. It is expected that within half a year, we will be able to use domestically produced photoresists." The photoresist is gone," Zhang Rujin said.
"Within half a year? Okay, then we will increase the quantity of stock. Even if we no longer import photoresist, it will still take half a year." Qin Tao said.
In fact, the more advanced the process, the better the quality of the photoresist required. When the technology of a few nanometers or tens of nanometers is used in later generations, the photoresist required is the most advanced. The current 0.13 micron technology will not be able to be used in future generations. The low-end photoresist can be produced by oneself.
“Mr. Qin, are you worried about others stumbling upon us?”
"I think it's not a worry, but a certainty." Qin Tao said: "We used to produce our own products, which made people hate us. Now, we have to do OEM work, which makes many people's teeth itch. They first talked to us. Sign the contract, and when we suddenly encounter supply cuts and no production, they can come and seek compensation."
"So this time is a conspiracy by AMD?" Zhang Rujin became nervous. He was a technical person and was really unfamiliar with such political matters. Thanks to Mr. Qin's reminder, otherwise he would have fallen into a trap.
"I don't know about this. Maybe they really want to find us as an OEM." Qin Tao said, "However, other people will take this opportunity to create crises. However, we don't have to worry about it. We just need to do our own thing well." good."
“Actually, we don’t have to do OEM work…”
"Of course we need OEM." Qin Tao said affirmatively: "Only OEM can allow us to grow and develop at SMIC. We don't engage in design. If we don't do OEM, how much money will we make? How can we grow and develop without making money?" ? We not only want to OEM for domestic chip design units, we also want to OEM for other chip design units in the world. In the future, we will squeeze out all foundries and monopolize this market."
This monopoly is not a brand monopoly, it is just a monopoly of the foundry. Others cannot find fault. Once oneself is in a monopoly position, it means that it has the right to speak. There will be no shortage of struggles in the future. If oneself controls this monopoly position, it will There will be one more trump card.
A certain country wants to impose a fine if it wants to have a monopoly? Damn it, I won’t do OEM work for you, and your chips will go back several generations!
Any antitrust talk is nonsense. Boeing merged with McDonnell Douglas, leaving only one Boeing in the U.S. market. Why don’t we see Americans jumping out and calling for monopoly? Those military industry giants have continued to merge and have long had a monopoly.
Zhang Rujin also became excited. Mr. Qin made it very clear that he wanted to be the leader in the field of contemporary manufacturing and squeeze TSMC out of business!
Zhang Rujin and TSMC are old enemies!
Zhang Rujin’s heart became hot.
This night, Mayer, who came from across the ocean, didn't sleep well either. He was in the senior presidential suite of the Huating Hotel, looking at the brightly lit big city below, which gave him a strange feeling.
The East is taking off!
The East is like a sleeping lion, now it has woken up! It is shouting to the world that it wants to return to the top!
Once our chip business is handed over to Dongfang OEM, can we still regain control of chip production?
No matter what, the most important thing now is that AMD can survive!
After holding back for so many years, the K7 carefully built by AMD will definitely surpass Intel. Now, with Eastern technology, they can quickly surpass Intel and defeat Intel first.
Anyway, when it gets to the user, it will have the AMD trademark on it, and the name of the foundry will definitely not be written on it.
It wasn’t until late at night that Mayer finally fell asleep. In his sleep, he dreamed of AMD becoming the number one chip maker in the United States.
The next day, when he walked into the conference room and saw the people inside, he was immediately surprised: "Mr. Qin, you actually came in person?"
(End of this chapter)