Nvidia patents chip trick to save power by slowing down data processing
Nvidia describes a method for computer chips to save energy by identifying specific types of data packets that require less processing power. When the chip detects these 'low-power' packets, it temporarily turns off parts of the circuitry that are not needed for that specific task. This allows the hardware to consume less electricity while still handling the incoming data.
In the filing’s own words
A device includes receiver logic to receive frames over a link, datalink logic coupled to the receiver logic to process the frames, and control logic coupled to the receiver logic and the datalink logic. The control logic is to determine whether a first frame indicates that a first subsequent frame is a low-power frame type for which the datalink logic consumes less power than when processing a second frame type, and to cause the datalink logic to process the first subsequent frame as the low-power frame type by enabling a clock gate for at least a portion of the datalink logic used to process the low-power frame type, responsive to determining that the first frame indicates that the first subsequent frame is the low-power frame type.
What exactly they claim (claim 1)
1. A device comprising: receiver logic to receive a series of frames over a link; datalink logic coupled to the receiver logic, to process the series of frames; and control logic coupled to the receiver logic and the datalink logic, the control logic to: determine whether a first frame indicates that a second frame is a low-power frame for which the datalink logic consumes less power than when processing a client frame; and responsive to determining that the first frame indicates that the second frame is the low-power frame, enable a clock gate for at least a portion of the datalink logic used to process the low-power frame.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- LOW-POWER FRAME TRANSMISSION OVER A COMMUNICATION INTERCONNECT
- Publication number
- US20260277844A1
- Application number
- 19682207
- Filed
- May 19, 2026
- Published
- Sep 17, 2026
- Inventors
- 2
- Claims
- 20
Classified under
G06F 13/4068 Physics & computingG06F 2213/40 Physics & computingMore from Nvidia in Chips & hardware
- Nvidia patents method to skip zero data in neural network math Oct 13, 2026
- Nvidia patents hardware to speed up motion blur in ray tracing Oct 13, 2026
- Nvidia patents method to sync data across split processor chips Sep 17, 2026
- Nvidia patents chip math trick to speed up AI calculations Oct 8, 2026
- Nvidia patents data center cooling that uses cryogenic fluids Oct 15, 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.