How will cloud phone application scenarios be innovated in the AI era? The new generation of intelligent cloud Panyu hive server tells you the answer!
Recently,The new generation of intelligent cloud Panyu hive server was releasedThis is a SoC array server developed based on the new generation QCS8550 chip, which has the characteristics of high performance, high-density computing power, low latency, and high energy efficiency. This model of server has been officially launched to the market, and will be widely used in cloud gaming, digital human live broadcast, AI large model training, cloud marketing, cloud office and other scenarios.
The appearance of the new generation of Panyu hive server.
Compared to the previous generation of Panyu Hive server,The new generation of Panyu Hive server has increased the single-chip CPU computing power by 50% and the GPU computing power by more than 230%, which can be called the "king of cost performance"!
The new generation of Panyu Hive server is equipped with QCS8550 chip, which uses a 4nm process and has a super-large core Kryo Prime, up to 32GHz frequency, with excellent single-thread performance processing capability; It also features four performance cores and three efficiency cores, which also excel in multi-threading. At the same time, up to 72 QCS8550 chips are integrated in a server chassis, and the whole machine has up to 48TOPS*72 NPU computing power, and the AI computing performance is also very strong, which can meet the business needs of customers with high density, high concurrency, and high computing power.
Specifically, the new generation of Panyu Hive Server will bring many new possibilities for cloud phone applications with four major advantages:
>> is more cost-effective, helping enterprises reduce costs and increase efficiency
How to realize the transformation from "electric tiger" to "digital economy engine"? Based on the intelligent chip with large computing power and low power consumption, the energy efficiency ratio of the new generation of Panyu Hive server is more than 2 times that of the traditional x86 server when completing the same computing tasks, but the cost is reduced by more than 50%, which can effectively help the data center obtain a cost-effective computing power solution!
>> Enhanced rendering capabilities for a better cloud gaming experience
How can popular games make players' experience smoother and more stable? It is also inseparable from the new generation of Panyu hive server! In terms of rendering capabilities, the GPU rendering capability of the QCS8550 chip can better cope with tasks such as image processing and graphics rendering, support ray tracing technology, and achieve tens of billions of ray interactions per second, ensuring that high-quality rendering services can be continuously and stably provided in various complex scenes.
For example, in a cloud gaming scenario, the Pantama Hive server can bring players to lifelike light, reflections, and lighting effects, and provide a smooth, realistic visual experience. For example, in the recently popular "Genshin Impact" 3D game performance test, the Panyu Hive server performed very well, providing a smooth and stable gaming experience; In terms of image quality, the strong GPU rendering capability of the Panyu Hive server makes the lake ripples in the game very realistic, and the realistic picture effect can be comparable to the real flagship machines on the market.
Gaming tests were rated at 60fps
>> The image processing capability is more efficient, and the image of the digital human is more realistic
When you swipe the live broadcast room every day to buy, buy, buy, buy In scenarios that require high image processing capabilities, such as digital human live broadcast, the new generation of Panyu hive servers is also helping.
With its powerful and efficient image processing capabilities, Panyu Hive Server enables it to quickly process various complex operations in digital human live broadcast, so as to ensure the smoothness and real-time performance of live broadcast. The digital human image is also able to present realistic visual effects, which are more vivid and natural.
>> AI computing power is more efficient and the deployment cost is lower
How to make full use of computing power has become the key for enterprises to control costs and improve R&D efficiency.
In terms of AI computing power, the new generation of Panyu Hive server is equipped with a newly upgraded Hexagon processor, which supports micro-slice inference and a tensor accelerator with stronger performance, and the Intel8 computing power is up to 48TOPS, which can support a variety of AI algorithms such as face recognition, voice activation, and eye detection. According to the performance evaluation of large language models, the performance evaluation of the new generation Panyu Hive server based on the LLAMA 2-7B large model can reach more than 20 tokens, which can meet the needs of large model inference and provide cost-effective AI performance support for enterprises and developers at a very low cost.
Based on the Panyu Hive server with independent intellectual property rights and innovative containerization technology, ARM cloud products continue to build a rich ARM cloud product matrix with the value concept of "high security, high energy efficiency ratio, and high cost performance", which can be applied to cloud gaming, digital human live broadcast, AI large model training, cloud marketing and other scenarios, helping customers obtain safe, stable, green and energy-saving, efficient and agile ARM cloud services and cloud computing power at a lower cost, and creating more possibilities for the digital transformation and intelligent upgrading of enterprises.