Reborn in 2011, 985 missing from two copies

Williams is responsible for the development of two mobile phone chips at the same time, and he has been under a lot of pressure recently:

"Chairman, the 40-nanometer Kunpeng 500 is a system-level chip. Since it integrates the Yilong 2800 baseband, it is more difficult to develop. It is better to hand over the 28-nanometer Kunpeng 700 project and team without an integrated baseband to George, and I will assist him."

"This way I can concentrate on solving the baseband integration problem of Kunpeng 500. As a result, Kunpeng 500 is expected to be taped out ahead of schedule, almost in October."

"George's team took over the Kunpeng 700 chip without a baseband. With our Kunpeng 500 technology and experience, the 700 will be smoother. In addition, without a baseband, even at 28 nanometers, there is hope that it will be taped out in advance, and it is estimated that it will be around March next year."

Theoretically, the 28nm 700 is more difficult to develop than the 40nm 500.

But the most difficult part of developing mobile phone chips is not the process, but the integrated baseband!

The Kunpeng 700 is an external baseband and has no integrated baseband at all. It only needs to upgrade the CPU and GPU based on the Kunpeng 500.

There is no need to even develop a baseband. You can just take the baseband from Kunpeng 500 and plug it into 700, thus saving yourself a huge problem.

For Kunpeng 500, not only the CPU and GPU need to be developed, but also the baseband needs to be successfully integrated! This greatly increases the difficulty, and is even more difficult than the 28-nanometer 700.

In addition, the development time of Kunpeng 500 SOC was even tighter, so Williams took on the most difficult part himself. Wang Yi had no objection:

"Okay, the Kunpeng 700 chip without integrated baseband will be handed over to George's team. William, you focus on getting the 500 SOC done! Get it done as soon as possible!"

As a result, Williams is responsible for the 40nm Kunpeng 500 SOC with integrated baseband.

The 28nm Kunpeng 700 chip project with external baseband was handed over to George.

Two teams, two projects, advancing simultaneously, supporting and assisting each other.

After all, Kunpeng 500 and 700 are both based on the same architecture, but the 700 has a higher main frequency and a more advanced process. It can be regarded as a semi-upgraded version of 500 without the baseband.

Once Williams has finished developing the Kunpeng 500 SOC, it will directly develop the next-generation 28nm Kunpeng 900 SOC with integrated baseband!

The two teams help each other and compete internally, which greatly improves efficiency and makes them more reliable.

What’s more, as Xingyi Technology develops further and enters the PC industry in the future, it may not necessarily develop its own PC chips!

For the chip R&D department, it would naturally be better to have two more teams!

Williams and George are both teachers and friends. This relationship is very similar to that of Mr. Hu and Mr. Liang, who are also both teachers and friends.

Rather than being wary of each other or harming each other, Wang Yi appreciates this kind of demeanor of being both a teacher and a friend, and helping each other.

"No problem!" Williams responded readily.

"That's fine with me too." George was even more motivated.

Wang Yi was in a good mood: "Then I will wait for your good news. Now that the R&D plant of Xingyi Wafer Factory has been delivered, it is expected that the 40-nanometer process can be trial-produced in August. If everything goes well, 40-nanometer can be mass-produced around October. If the Kunpeng 500 soc can be successfully developed ahead of schedule, it will be ready for tape-out at that time!"

"Okay, Chairman, we'll do our best!" Williams promised.

Wang Yi nodded and left the chip department.

鲲鹏500 SOC，40nm，集成40纳米的翼龙2800基带，性能比45纳米的tegra 3略强。

But it is not as good as Qualcomm's 28nm APQ 8064, also known as Snapdragon 600. This year's absolute flagship, next year's sub-flagship chip.

Nvidia's flagship tegra 3 this year is only equivalent to next year's Qualcomm Snapdragon 400!

After all, the Tegra 3 is 45 nanometers, and Qualcomm Snapdragon 400 is 28 nanometers.

The CPU of Snapdragon 400 is not as good as that of Tegra 3, but its GPU, power consumption control and baseband all surpass Tegra 3.

In this way, the overall performance of Kunpeng 500 is slightly better than next year's Snapdragon 400, but not as good as Snapdragon 600.

This year, Kunpeng 500 can be positioned as a sub-flagship chip, used to replace the NVIDIA quad-core tegra 1 in Xphone 1pro, Xingyi tablet X3pro, and Xingyi TV.

Next year, it can only be positioned as a mid-range chip, suppressing Snapdragon 400, and used in borderless thousand-yuan phones.

The chip upgrade process is so fast, one level lower every year.

As for Snapdragon 600, it is suppressed by the 28nm Kunpeng 700!

Kunpeng 700 does not have an integrated baseband, so it cannot compete with Snapdragon 800, but it can easily crush Snapdragon 600.

In addition, although Kunpeng 700 does not have an integrated baseband, it has powerful performance and supports 4K decoding, so it can be used in next year's high-end Xingyi TV 4K version!

Next year's Xingyi tablet X2pro can also be upgraded to Kunpeng 700.

As for next year's Qualcomm flagship SOC Snapdragon 800, we can only wait and see the progress of Kunpeng 900.

Although they are one step slower than Qualcomm, there is no other way. After all, Xingyi Semiconductor is a rising star, and it is already very good to be able to achieve this step.

Fortunately, Texas Instruments gave up its mobile chip business, and Wang Yi directly took over a large number of Texas Instruments' R&D engineers, semiconductor gurus, and experts. It was equivalent to standing on the shoulders of Texas Instruments to develop, and naturally the development was much faster.

If Xingyi Semiconductor had not made advance arrangements and recruited the big names and R&D engineering team from Texas Instruments, it would have taken an unknown amount of time for it to develop.

If we develop it ourselves, at least the first generation of chips will most likely be like Huawei's K3v2, and it will be difficult to achieve success.

Wang Yi took a shortcut. He not only took over the CPU and GPU R&D experts and team of Texas Instruments, but also incorporated VIA's baseband R&D team and 55Nm baseband technology.

The reason why the 40-nanometer Wing Dragon 2800 baseband has made rapid progress is that it was developed based on VIA's 55-nanometer baseband technology.

With the technical and R&D experts from the two teams, plus Wang Yi's strong financial support and active recruitment, the progress was rapid!

It can only be said that Wang Yi had the best of both worlds.

It just so happened that Texas Instruments was going to give up on mobile chip research and development, and it just so happened that VIA was also going to give up on baseband research and development, and in the end, both benefited Wang Yi!

This is the biggest advantage of the reborn, control over information and continuous picking up of leaks!

The next day, Baidu headquarters.

Li Yan looked at the sales of Xiao Ai speakers and was very envious:

"Wang Yi is really amazing. He made a big splash with a small speaker, and even added AI and smart home features! It's selling so well, it's really amazing."

The vice president also had a complicated expression: "Especially with the infrared function added to Xiao Ai Speaker, it is simply invincible. Originally, those traditional air conditioners and TVs could not be controlled by Xiao Ai Speaker, but with the infrared function, they can also be controlled directly."

Another executive said: "The most impressive thing is the layout of smart home! I feel this is a new track with great prospects."

Li Yan agrees: "Well said. In the social field, Baidu cannot compete with Tencent's QQ or Xingyi Technology's Changliao. In the video field, iQiyi is no match for Xingyi Video. But in the field of smart home, we have great potential!"

The vice president nodded and said, "Yes, Xingyi Technology has advantages in AI speakers, and so does Baidu. Xingyi Technology has Xingyi Music to provide copyrighted content, and we at Baidu have Qianqian Jingting, which also has genuine content. Although our music library only has millions, not as good as Xingyi Music's tens of millions, it is enough."

"Besides, Baidu already has an advantage in the search field. Once we launch Baidu AI speakers, our human-computer interaction and question-answering will be more powerful than Xiao Ai speakers. Similarly, we have accumulated some technologies in AI. As long as we invest money and strengthen research and development, we can make Baidu AI speakers in a year at most!"