Snapdragon X80 5G Modem – RF Platform
Qualcomm
Much has been said over the past decade about Moore’s Law. Most of them have to do with the system being down or at least slow. Clearly, breakthroughs are still being made in smaller process node geometries by leading companies such as Intel and TSMC. But it is equally clear that the industry can no longer rely solely on Moore’s Law to improve performance. Some companies are trying to achieve higher performance using heterogeneous computing architectures, some are using chiplets, and some are using both. At this year’s Mobile World Congress, Qualcomm highlighted one more tool that can be used, along with its established heterogeneous computing architecture approach: artificial intelligence (AI).
A popular saying goes something like, “If something is great and you want it to be even better, add more bacon.” Whether you agree with this or not, or just replace “bacon” with something else, the sentiment that there are certain things that make things better if you add them still holds true. AI is becoming just that for the technology industry, especially the chip industry.
This isn’t just about generative AI. Generative AI has matured enough over the past few years to have a significant impact. In contrast, adding traditional or functional machine learning-based AI to enhance product functionality has been a differentiation strategy for the past decade or so. At his MWC this year, Qualcomm applied this potential success recipe to his 5G and unveiled its latest 5G product, his Snapdragon X80.
AI as the main character
Snapdragon X80 is a modem RF platform that consists of four components: baseband, RF transceiver, RF front end, and mmWave front end module. Supporting 3GPP Release 17 and expected Release 18 features, the X80 modem supports 6x downlink carrier aggregation, up to 6 receive channels for smartphones, 10 Gbps peak download speed, 3.5 Gbps peak upload speed, narrowband non-terrestrial Enables support for wave networks. 5G optimization with AI, of course.
The baseband is equipped with Qualcomm 2n.d. Generation 5G AI processors enable platforms to leverage AI to improve quality of service and end-user experience by intelligently controlling both modem and RF functions. Together with their three,rd With the fifth generation 5G AI suite, performance metrics such as data speeds, power handling and efficiency, coverage, spectral efficiency, latency, and GNSS location will all be improved through the use of AI. Additionally, AI processing is used to assist with mmWave beam management. This is essential to provide 5G mmWave range extension when used in fixed wireless access (FWA) customer premise equipment (CPE).
These improvements are achieved in part by leveraging AI to more intelligently and efficiently manage multi-antenna subsystems. AI is also utilized to provide contextual input and optimize radio links by identifying and considering the state of the RF environment and what the user is doing in terms of the application or workload. For example, if a user is doing something latency-sensitive like a video call, AI could increase transmit power to compensate for the decrease in channel quality, prioritizing throughput and latency over the resulting increase in power consumption. There is a gender.
Compared to the previous generation, Qualcomm says it implements AI to optimize 5G performance, reducing best cell selection time by 20%, link acquisition by up to 30%, and same location accuracy. can be improved to a certain extent. For mmWave applications, CPE service acquisition is up to 60% faster and consumes 10% less power during the connection.
Use AI to make it better
As mentioned earlier, Qualcomm wants to dramatically improve the user experience with its latest X80 modem RF platform. While this is always the goal with each new generation, advanced use cases and applications involving generative AI increasingly require faster processing, faster throughput, and latency while preserving or increasing power. It is needed now more than ever. consumption on these devices.
In addition to throughput and latency, user experience also depends on battery life. Tirias Research recently conducted a study on the latest flagship smartphones running generative AI workloads and determined that current battery technology needs all the help it can get in the AI era.
Applying AI to ensure the best combination of modulation coding scheme, transmit power, and antenna array configuration for a given workload can help in many ways. AI is already a key factor in improving the user experience by maximizing uplink and downlink throughput and minimizing latency, as well as improving the transmission chain (RF front end RF transceivers and power power-on time (amplifier) and power-up time to a minimum. This ensures that the power used for transmission is as strong as necessary to achieve the desired performance. The transmission chain, along with the application processor and display, is one of the biggest consumers of power provided by a mobile device’s battery. Extend battery life by minimizing the time the transmit chain is on, either by minimizing retransmissions to overcome high error rates or by maximizing throughput and minimizing delay This can have a huge impact on the user experience.
AI-optimized experiences don’t just benefit end users. If a given transmission can be completed faster with lower transmit power, mobile network operators can also maximize capacity through a lower effective noise floor, ultimately minimizing interference and achieving the highest possible downlink and uplink speeds are achieved, benefiting end users. For certain RF environments.
What’s next for 5G and AI?
In addition to AI-based enhancements, the X80 Modem RF platform also supports anticipated 3GPP Release 18 (Rel 18) features. That’s why Qualcomm claims that this latest of his 5G products is “5G Advanced Ready”. Regardless of which features end up being standardized in Release 18, OEMs that design at least X80 modem RF into their device lineups will benefit from the next generation of not just smartphones but other types of devices as well. You’re getting what Qualcomm thinks will be value-added features. Same goes for PCs, XR devices, cars, FWA CPE, etc. Qualcomm says the platform is currently in the sample phase and is expected to be commercially available later this year.
Qualcomm’s use of AI to improve the performance of its products is not a new idea, and it is not the first time Qualcomm or other companies have done this. Notably, this product announcement reflects the important role that traditional AI plays in enhancing existing technologies and products to a level that not only enables but also uses workloads such as generative AI. , and serves to highlight the important role it will continue to play. It also demonstrates another weapon in the chipmaker’s arsenal to continue delivering the needed performance gains, with or without Moore’s Law.
Qualcomm isn’t the only company leveraging AI in this way. However, in a world focused on generative AI, it is easy to forget the importance of traditional machine learning-based AI product optimization. Traditional AI used in this way is a powerful tool for differentiation, and every opportunity to leverage it should be taken advantage of.
follow me twitter Or LinkedIn. check out my website.
The authors and members of Tirias Research staff hold no shares in any of the companies mentioned. Tirias Research tracks and consults companies across the electronics ecosystem, from semiconductors to systems and sensors to the cloud. Members of the Tirias Research team are tracking all developments in 5G modems and AI and consulting for Qualcomm, Nvidia, MediaTek, and other companies focused on cellular connectivity and AI.
