Nvidia patents parallel method to assign wireless frequencies to devices
Nvidia describes a method for using multiple processor threads to quickly assign specific frequency bands to wireless devices. The system runs a peak-finding algorithm in parallel to determine which devices get which parts of the spectrum based on performance metrics.
In the filing’s own words
In various examples, apparatuses, systems, software, and methods are described for executing a Riding Peaks algorithm in parallel using multiple threads (e.g., GPU threads) to allocate sub-bands to user equipment device (UE)s. Examples of UEs include mobile phones, in-machine infotainment devices, smart devices, security systems, and/or other wireless-capable devices. For example, a processor executing an algorithm that allocates sub-bands to UEs by using threads (e.g., GPU threads) operating concurrently to identify peaks and assign sub-bands.
What exactly they claim (claim 1)
1. One or more processors comprising: circuitry to: receive performance metric values for a plurality of user equipment devices (UEs); identify one or more first peaks based at least on determining that one or more performance metric values are: (i) greater in a sub-band compared to other sub-bands for a UE of the plurality of UEs; and (ii) greater compared to the same performance metric for another UE in the sub-band, the determination performed in parallel by executing a first reduction operation using first threads and a second reduction operation using second threads of at least one processor of the one or more processors; identify one or more second peaks based at least on performing a parallel prefix minimum operation, the parallel prefix minimum operation performed using third threads and fourth threads of the at least one processor in parallel to compute prefix minimum values across sub-bands; and allocate one or more sub-bands to the UE of the plurality of UEs based at least on the identified one or more first and one or more second peaks and gaps between assigned sub-bands based at least on executing parallel threads for individual gaps of the gaps.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- PERFORMING FREQUENCY ALLOCATION IN PARALLEL
- Publication number
- US20260312700A1
- Application number
- 19194373
- Filed
- Apr 30, 2025
- Published
- Oct 8, 2026
- Inventors
- 2
- Claims
- 20
Classified under
H04B 17/309 Electricity & electronicsG06F 9/30185 Physics & computingG06F 9/3888 Physics & computingG06T 1/20 Physics & computingMore from Nvidia in Networks & wireless
- Nvidia patents AI method to pick best wireless signal beams Oct 6, 2026
- Nvidia patents method to count active wireless devices in a network Oct 6, 2026
- Nvidia patents using neural networks to simplify 5G signal data Oct 13, 2026
- Nvidia patents using AI to prioritize phone connections Sep 10, 2026
- Nvidia patents method to keep cloud video and audio in sync Sep 17, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.