"Duke, our mobile CPU chips are now making great progress. Following Gemini2, our 4-core products Gemini4 and octa-core products Gemini8 are now in mass production. We have begun to provide these chips to SAL mobile phones, TVs and smart car electronics, which can fully meet the requirements of SAL's embedded system chips. The powerful gemini8 can even replace other companies' high-performance desktop-level CPUs and start to be used on advanced CNC machine tools and nuclear power plant industrial control systems.--) Next, we will start developing native 12-core products and superimposed 16-core products."

Williams reported on his progress during this period. Duke has read the report and is familiar with the situations he said. After breaking through the Gemini2 design, Duke's investment in NDK increased greatly. Therefore, when Williams developed Gemini4 and Gemini8, he adopted parallel research.

Relatively speaking, there is not much essential difference in technical difficulty between 4-core and 8-core. Williams has more than doubled the R&D team and invested it in the two teams. Research progress with sufficient resources is very smooth. In addition, Williams himself is very confident in these technologies. These technologies are not the most advanced today. The competitor's Arm series has already completed the 8-core design and is now entering the 16-core era.

It is just that Duke has been optimized in the design of gemini2 architecture, and these new technologies are also applied to gemini4 and gemini8.--(Because whether it is gemini4 or gemini8, Williams adopts the native 4-core and 8-core design method, and does not superimpose two gemini2 cores to become gemini4. Although such a design is much more complex, the system performance is one level higher than that of the same core chip of the arm. So in fact, the gemini8 developed by Williams has far exceeded the performance level of the arm8 core.

Now Duke has designed a large number of solutions to adopt gemini8 in various places where high-performance chips are needed, including on the J7 drone prototype. Duke has used hundreds of gemini8s, so Duke is not unfamiliar with these two chips.

Williams naturally did not come to develop the gemini16 core product. According to the plan, the gemini12 is native 12 cores, and the gemini16 core product will use two gemini8 cores to superimpose it. This situation is not particularly important news. The relevant plan has long been approved by senior Sal R&D leaders such as Zhao Jianwu.

So Williams is just an opening remark, "Okay, you all know these things, and I'm not here for this today," Williams looked at the two young men in front of him calmly, waiting to say the key points, and shrugged and changed the topic. In fact, he is not a good lobbyist, so the effect of these preparatory things seems to be very failed.
Chapter completed!