As we all know, the content that smart phones need to deal with is becoming increasingly complex. Gone are the days of the first generation of single, monochrome display screens. Today, even mainstream and entry-level mobile devices need to support a rich multi-layer user interface and many of the latest applications and technologies. These complex user requirements continue to push mainstream mobile devices to higher performance. For millions or even billions of mobile end users around the world, being able to support these functions and uses is critical.
Therefore, Arm recently announced the launch of two new Mali GPUs: Mali-G52 and Mali-G31, providing solutions for mainstream and ultra-high-performance devices. With a new execution engine design, Mali-G52 can achieve higher performance in a smaller chip area, supporting mainstream devices to achieve advanced functions such as machine learning (ML) and augmented reality (AR); while Mali-G31 is Arm to date The smallest GPU introduced so far can support the latest generation of APIs for digital TV (DTV) graphics development and complex user interfaces. So, how can these two new GPUs help the next generation of new technology?
Machine learning is becoming more and more indispensable
Maybe you think that only the most high-end mobile devices will be equipped with machine learning functions, but what is surprising is that everything is developing at a speed that exceeds your expectations. Even on mainstream mobile phones, we hope that the camera of the device can automatically recognize the face in order to focus accurately, or be able to accurately search for photos of pets in the album. You may not realize that these functions are all realized by machine learning. Arm recognized the importance of machine learning and recently released the Project Trillium machine learning platform in a targeted manner. However, although Arm's powerful new-generation machine learning processors perform very well for high-end machine learning workloads, this is only one of the more widely used scenarios. As we discussed recently, in mainstream and entry-level devices, it is often impractical to equip such a processor with a sufficiently powerful chip, but for functions with less load like those I mentioned before, we A GPU or CPU is usually suitable. We have previously discussed how our GPUs are handy for such workloads, and also talked about how our latest generation of Arm Cortex CPUs achieve specific machine learning gains.
Our newly released Mali-G52 has pushed all performances up one level again. On top of the success of the previous generation of GPUs based on the Bifrost architecture, Mali-G52 uses a newly designed execution engine-the part of the processor that executes the algorithm, thereby achieving higher requirements for machine learning within a lower chip budget performance. Compared with the previous generation processor, Mali-G52 doubles the number of each execution engine thread, which doubles the calculation performance of all complex content, but the size is only increased by 22%, so it will not worry partners. Chip budget. Not only that, Mali-G52 also introduced Int8 dot support. The machine learning derivation algorithm on the terminal widely uses general-purpose matrix multipliers, but usually does not need to reach the accuracy level of FP16 / FP32. In many cases, Int8 is equivalent and more efficient. The processing power of each execution engine of Mali-G52 reaches four cycles per thread, and it achieves an excellent performance of nearly 4 times the machine learning performance of the previous generation processor in the current image detection and other machine learning benchmark tests.
Our partners also attach great importance and affirmation to this performance. Daniel Diao, deputy general manager of Huawei's Turing Processor Division, said: "As user demand grows, machine learning has become the most important element of all our devices. This GPU can be of such excellent quality Performing challenging machine learning task loads, this performance will completely disrupt the mainstream end market. "
Mixed reality will prevail
The more immersive content experience will no longer be exclusive to the most high-end devices. When you do n’t want to wear a VR headset for a full hour (maybe you do n’t want to wear any additional equipment at all, but you can still enjoy a fully immersive experience in some delicate ways), it does not mean that there is no other implementation Selection of VR effects. AR augmented reality is quickly becoming the first choice of reality technology, and the ability to support this function and 360-degree panoramic video in mainstream devices has become critical. With the highest energy efficiency in the same series, the Mali-G52, together with the intelligent task distribution function of the DynamIQ CPU architecture, has truly pushed its performance to the peak. In this generation of mainstream mobile solutions, the most respected combination is a DynamIQ configuration consisting of one Cortex-A75 processor and seven Cortex-A55 processors. These smaller cores can simultaneously handle multiple tasks that are less tedious, This allows the powerful Cortex-A75 to focus on areas where it really needs to exert its overall performance. Thanks to the above progress, the performance density of Mali-G52 is 30% higher than the previous generation, and it can handle highly complex content in an instant. Similar to the flexibility of DynamIQ, Mali-G52 has extremely scalable implementation options, allowing our partners to accurately achieve the balance of performance and efficiency required by their equipment. Partners can choose two or three execution engines for each core and up to four cores. This additional scalability has significant advantages for low-cost mid-range and low-end devices because it provides each partner with a very high level of granularity when weighing priorities.
3D games everyone loves
In the past, if your support on your phone was below a certain level, you would n’t think there would be a better choice than a gaming experience like Snake Snake. Today, as the display is more effective and the visual effects are more brilliant, we have also raised expectations in terms of user experience. Last year we talked a lot about the content of the mobile game industry surpassing console games and table games in terms of revenue and attractiveness driven by high-fidelity mobile games. This trend also extends to the mainstream and ultra-efficient terminal fields. This is our first Two new GPUs-Mali-G31. To meet the needs of developers, Mali-G31 is Arm's first ultra-efficient GPU based on the Bifrost architecture. Mali-G31 is the smallest processor that can support both OpenGL ES 3.2 and Khronos API Vulkan. This means that developers can access these APIs through a new content performance layer to achieve huge energy efficiency and optimization. Currently, millions of mainstream and entry-level devices are active on the market, and application developers can optimize their content to achieve perfect operation on all levels of devices driven by the Mali chipset. This brings not only quality improvement, but also performance in terms of quantity-more games and applications will run on more devices, thereby benefiting more users. This also means that developers can enter a larger market and continue to develop the next stage of the application. From silicon chip partners to game players, all links in the entire mobile ecological chain will benefit from this.
Embrace the TV revolution
We have talked about the status and future of these brand new technologies, but these technologies are not the only factors pushing today's low-cost systems to their limits. If you have only recently upgraded to a smart TV, then in terms of user interface and interactivity, the new chip technology will lead a huge leap in user experience compared to the previous generation chip technology. Today, we no longer just watch a group of channels from a content provider, or watch a group of channels at once. Since streaming services such as Amazon Prime and Netflix have been directly integrated into the system, the system must support many different interfaces, graphics overlays, and display composition layers. In addition, many users now use voice-activated search to select the programs they want to watch. The seemingly simple digital TV has developed to a very intelligent level. Mali-G31 will enable the future of digital TVs by reducing its size by 20%, as well as reducing the cost for chip manufacturers while also meeting the task load of user-specific interfaces. When the new mainstream solution is implemented together with other new Mali multimedia suite products-Mali-V52 and Mali-D51-, it will bring significant performance and efficiency improvements and real competitive advantages to our customers.
With the continuous development of innovative technologies, the development territory of Arm and partners is constantly expanding, and the mainstream device market is entering the domain that was originally exclusive to the high-end market, thus providing a new generation of consumer groups with the next generation of visual experience and engaging immersive content .
About Arm
As the core of the computing and interconnected revolution, Arm technology is changing the way people live and operate their businesses. From indispensable areas to invisible support, Arm's advanced energy-efficient processor design has been applied to more than 100 billion chips to safely provide support for electronic devices, covering a variety of applications from sensors to smartphones to supercomputing. Arm has more than 1,000 technology partners, including the world's most famous commercial and consumer brands. Arm is actively collaborating and expects to apply Arm innovation to all areas that require computing, including chips, networks and clouds.
BAM Power Transmission Capacitors
BAM power transmission capacitors
RAM/RFM Induction Heating Capacitors,Water Pump Capacitor,Water Cooled Condense,Tank Capacitor
YANGZHOU POSITIONING TECH CO., LTD. , https://www.cnchipmicro.com