When Matt got off the plane, Duke had no idea that this slightly melancholy blonde man was here for him, Duke who came back from the new ceremony, was unaware of these undercurrents against himself, and now he was almost fully engaged in the J-7 drone transformation project, for this project that Mr. Liu had high expectations, Duke happened to have a little free time on other projects, so he put a lot of effort into it, ready to do it well.

Compared with Jingwei and Pterodactyl, this transformation plan is very challenging, first of all, because the combat radius used by this UAV program far exceeds the current level of Jingwei and Pterodactyl-class UAVs, according to the combat concept of the final assembly, the combat radius of the transformed unmanned J-7 fighter is at least a few hundred kilometers or more, Duke needs to develop a new autonomous identification system for them to find and lock on combat targets.

Because this combat radius has exceeded the limit of the wireless communication technology developed by Duke, a large number of complex calculations can no longer be performed by transmitting real-time data to the ground monitoring workstation, but must be completed autonomously on the J-7 UAV, and the transmission of this real-time data is no problem with the wireless communication technology developed by Duke within a range of two or three hundred kilometers, but the greater the distance, there is nothing that can be done.

Although satellite communication can be transmitted to farther places, the bandwidth cannot support the needs of large group operations for J-7 UAVs, so this complex calculation can only be completed on J-7 UAVs.

However, from the cost and cost performance point of view, it is not cost-effective to equip each J-7 UAV with a high-performance workstation that can independently complete this computing task, this high-performance workstation alone costs up to millions of yuan, and the configuration is not much different from that of a small machine. And these will increase the already small carrying weight of the J-7 UAV.

For Duke, this is certainly not an ideal option, although this scheme is much simplified in terms of application implementation, and it can be achieved by basically moving the ground monitoring station computing system of the Pterodactyl system directly to the J-7 UAV. But this is not in line with Duke's identity as a tech genius, and it is difficult to make Duke himself satisfied with such a bulky and uncreative thing.

Therefore, it is not a good option to independently configure a high-performance workstation for the J-7 UAV.

Duke took out the functional requirements and operational requirements provided by Xiao Qianguang, and carefully studied the combat methods that the general assembly envisioned these UAVs might use in the future, Duke found that there is a very significant feature, that is, these J-7 UAVs are rarely used for individual combat missions, and they are not needed for reconnaissance, Duke will create better specialized unmanned reconnaissance aircraft for reconnaissance, and after decrypting and copying the materials, sensor systems and power units of the Global Hawk, Duke is confident that his R&D team will soon be able to build a better reconnaissance aircraft than the Global Hawk, and that there is no need for a clumsy guy like the J-7.

The biggest use of this J-7 UAV is still the use of group fighting tactics, that is, the use of large group combat tactics in one swarm. Faced with enemy high-value cluster targets, such as airfields, heavy weapons positions, and even imaginary combat targets such as large fleets.

In this mode of operation, Duke felt that it was possible to use the cluster computing method, just as he used the supersmall machine to achieve similar supercomputer functions through cluster computing, connecting multiple adjacent J-7 drones through wireless broadband, preparing to make the future J-7 drones have cluster computing capabilities.

In this case, in order to adapt to the group fighting tactics of the J-7 drone in the future, there is no need for Duke to choose a super workstation with 8 latest CPUs to do this computing core, although the computing power of this workstation is very strong, but the price is 4 to 5 times more expensive than 4 similar low-grade workstation servers, and the computing power has not increased to so many times the price.

If you choose 4 CPU workstation servers with low cost, you can get the computing speed of 8 CPU high-end super workstation servers through joint computing of more than 3 workstations.

According to the simulation test above Kerry, the joint computing speed of 2 low-end workstations can basically be close to more than 80% of the computing performance of the original high-end workstations, and when access to 3 can be exceeded, and the larger the cluster of workstations accessed at the same time, the better the computing performance, although this performance expansion is not unlimited, because on the one hand, the speed of wireless transmission is limited, and this joint cluster operation requires a large amount of data exchange, and when the size of the group expands to a certain extent, it will face this data transmission bottleneck.

On the other hand, the more servers that are connected at the same time, in fact, it means that there will be more images that need to be processed at the same time, when several planes are flying next to each other, a large part of the images they shoot are the same, and the image processing application developed by Kerry can automatically recognize the same images, greatly speeding up the same image recognition that neighboring aircraft "see", but in the case of cluster sorties, the images captured by aircraft that are far away do not have so much similarity, so the processing will not be too fast.

According to this feature, Kerry has developed a new cluster computing application, which can form a small computing cluster of servers on 3-5 adjacent aircraft through a wireless communication system based on the spatial coordinate position information of the aircraft cluster, which can perfectly process the images of the passing area in real time, identify possible potential attack targets in the area and share them with the entire cluster, or transmit the results back to the rear for confirmation through satellites.

This adds countless complexities to the software architecture, without the assistance of Kerry's supercomputer, Duke will definitely not be able to complete these super complex cluster algorithms, but after such a toss, the cost of each J-7 drone on the workstation server will be reduced to one-quarter to one-fifth of the original, that is to say, the transformation cost of a single J-7 will be reduced by hundreds of thousands of yuan, considering that the future transformation is calculated on the scale of more than thousands, Duke's change will directly save more than hundreds of millions of yuan.

Science and technology is the primary productive force in this research and development process is vividly embodied, and there is no loss of performance and function at all, of course, this is for the mode of cluster combat, if a single machine is dispatched, there is still a huge difference, which once again shows how big the difference in value between the human brain and Kerry's intelligent brain.

(Special thanks to the book friend "lanxue" for supporting 6666 Zongheng coins, this is the largest reward received by the book, and also thanks to the book friend "Who am I 4004" for their enthusiasm, I can fully feel your enthusiasm, and today I will add a special chapter at 23:58 to congratulate the book friend "lanxue"!) Talk about Xiaofei a little heart! ）