The global upsurge in artificial intelligence applications, including Google, Apple, Microsoft and Intel, in 2017, has launched the AI app market with the full range of acquisitions of AI start-ups. The recent hot topics in the PC industry are Intel and Super Micro (AMD), which have been fighting for years, unexpectedly announced that will be the platform for cooperation, coupled with the former micro-mapping technology business operator Raja Koduri to Intel, the industry believes that Intel, Ultra Micro Recent actions, not only to consolidate the competitive PC market, But in an attempt to fully enhance the AI combat force, the Force to prevent Nvidia in the field of AI continued to sit large. The industry rumors of Intel and Nvidia to terminate the GPU license, will be another cooperation with the triangular relationship, recently officially announced, Intel not only hand over the micro, grant authority, and the cooperation between the two sides closer than expected, the industry interpretation of Intel, micro-unprecedented cooperation, It was all due to the powerful war of Nvidia. Based on Intel's platform cooperation program, the eighth generation Kaby Lake processor, the second generation of high-frequency wide memory (HBM2), and the Ultra Micro Radeon GPU will be integrated through the EMIB (Embedded Multi-die Interconnect Bridge) packaging technology, The industry, including ASUS, Hewlett-Packard (HP), Lenovo, Dell (Dell) and Acer and other notebook computers (NB) industry has received the Intel Platform Engineering test samples, is now conducting performance testing and benefit assessment. In addition, the original shoulder of the ultra micro-mapping technology, the senior vice president and the structure of the long Koduri, has suddenly changed to Intel, because Koduri is a global mapping industry heavyweight big Coffee, good at PC visual computing experience, from content creation to GPU system architecture design, Koduri pushed the first integrated high frequency wide memory (HBM) graphics chip in the ultra tiny time, and used to be the director of graphics architecture at Apple, leading the Mac product line mapping system into a retinal computer screen (Retina Display). Despite the fact that the ultra micro and Intel cooperation, but also lost koduri this mapping field of general, the industry believes that the ultra micro should have evaluated the pros and cons of cooperation with Intel, not only to achieve considerable benefits, improve profitability, more importantly, in the face of recent years Nvidia in the AI field strong rise, In the minds of enemy friends, the future cooperation with Intel will be more profound. In fact, the current Ultra micro in the NB processor or graphics chip market, the global market share is far behind Intel and Nvidia, this time Intel, Super Micro willing to work together, mainly has a common enemy nvidia, the two sides tried to compete in the electricity and AI applications market, comprehensive force to hinder nvidia expansion. Nvidia's rapid expansion of the global competitive PC battlefield not only controls the development of the hardware industry, but is on an equal footing with Intel, impacting Intel and its future development, as well as the industry's competition ecology, the supply chain to see Intel with the Ultra Micro Platform cooperation, hoping to further contain Nvidia. However, Nvidia has full control over the supply and bargaining advantage of the graphics card in the competitive PC market and will definitely retaliate against Intel and the Super Micro Alliance. In addition, Nvidia has been relying on the GPU technology for many years to stand out in the AI chip opener. Nvidia CEO Huang previously said Nvidia's continuous improvement of GPU design, system architecture, compilers and algorithms, which only took 3 years to increase the training depth of neural network 50 times times faster than Moore's law proposed time, and in the next few years can increase the speed of 10 times times. At present, in every PC, server, supercomputer, or Amazon, IBM and Microsoft Cloud Environment, developers can find Nvidia GPU, almost all the major AI development architecture using NVIDIA GPU. Among them, the GeForce chip can be used for PCs, Tesla for clouds and supercomputers, Jetson for robots and drones, and Drive px for vehicles that use the same architecture to speed up the depth of learning. Intel CEO Brian Krzanich said, looking to the future, AI strategy will become an enterprise survival requirements, enterprises need to immediately establish AI strategy and implementation, in order to win the future. In recent years Intel has fully developed AI soft, the complete solution of the hardware and the construction of the ecological chain, but in the very critical GPU parallel computing capabilities, is still less than Nvidia, now by the new war Koduri join, and with the micro-cooperation, Intel is trying to in the new generation AI battle, no longer let the Nvidia special beauty in the former .
