# How Graphics Processors Switch Between Different Tasks Efficiently > A method for graphics chips to pause and swap between different programs or tasks without waiting for every single part of the processor to finish its current job. - **Patent:** US 7512773 - **Original title:** Context switching using halt sequencing protocol - **Owner:** Nvidia Corp - **Granted:** 2009 - **Status:** Public domain (expired) - **Times cited:** 6 - **Field:** consumer_electronics, semiconductors, ai_ml ## What it does This patent describes a way to pause a graphics processing pipeline so the computer can switch to a different task. Normally, a processor might have to wait for every single unit to finish its work before it can switch tasks, which is slow. This method sends a 'request-to-halt' signal to all units. Each unit reports back whether it is busy, idle, or has successfully paused. Once all units are either idle or paused, a 'stay-halted' signal locks them in place so their current state can be saved to memory and replaced with the state of a new task. ## What it does NOT cover - Does not cover general-purpose CPU context switching that does not involve a graphics-specific pipeline. - Does not cover systems that require every unit to be completely idle before initiating a switch. - Does not cover software-only task switching that lacks the specific hardware 'stay-halted' signal architecture. ## The clever bit Instead of waiting for a total system stop, it allows units to enter a 'halted' state—where they stop mid-task but save their progress—allowing the system to switch tasks even when some units are still technically in the middle of an operation. ## Real-world examples 1. Nvidia GeForce graphics cards 2. GPU-accelerated multitasking in modern operating systems 3. Graphics drivers managing multiple concurrent applications ## Why it matters Graphics processors (GPUs) are designed to do many small tasks in parallel. Before this, switching between these tasks was inefficient because one slow unit could hold up the entire pipeline. This patent helped enable modern multitasking on GPUs, allowing the computer to switch between different applications or processes without a noticeable stutter in graphics performance. ## Frequently asked questions ### What does How Graphics Processors Switch Between Different Tasks Efficiently cover? A method for graphics chips to pause and swap between different programs or tasks without waiting for every single part of the processor to finish its current job. ### Who owns patent US 7512773? Nvidia Corp owns this patent, granted in 2009. ### When does this patent expire? This patent has expired and is now in the public domain — anyone can use the invention freely. ### What is patent US 7512773 cited by? This patent has been cited by 6 later patents that build on its ideas. ### What problem does this patent solve? Graphics processors (GPUs) are designed to do many small tasks in parallel. Before this, switching between these tasks was inefficient because one slow unit could hold up the entire pipeline. This patent helped enable modern multitasking on GPUs, allowing the computer to switch between different applications or processes without a noticeable stutter in graphics performance. ### What does this patent NOT cover? Does not cover general-purpose CPU context switching that does not involve a graphics-specific pipeline. **Full plain-English explainer:** https://patentbrief.org/patent/us/7512773/context-switching-using-halt-sequencing-protocol **Original patent:** https://patents.google.com/patent/US7512773 --- _Source: PatentBrief — https://patentbrief.org. Patent facts are from public records; the plain-English explanation is PatentBrief's._ ## Related patents Semantically similar inventions in the PatentBrief corpus: - [How Mobile Devices Switch Between Open Apps Using Gestures](https://patentbrief.org/patent/us/8291344/ios-autocorrect-keyboard) — Apple's patent describes a way to switch between open apps on a touchscreen by showing a bar of app previews and selecting one with a tap. - [How Intel's Memory Hub Manages Graphics Data Across Different Memory Types](https://patentbrief.org/patent/us/6362826/method-and-apparatus-for-implementing-dynamic-display-memory) — A hardware design for a computer memory hub that lets a processor treat different types of memory as one unified space for graphics tasks. - [How to Split Computer Tasks Between Different Types of Processors](https://patentbrief.org/patent/us/9477524/aws-cloudformation) — A system that uses a physical backplane to connect two different types of computer processors—one for general tasks and one for real-time tasks—to improve efficiency. - [How Salesforce Speeds Up Data Processing Using Bulk Message Handling](https://patentbrief.org/patent/us/10019297/power-bi) — A method for cloud systems to grab multiple tasks at once instead of one by one, using a locking mechanism to ensure data stays consistent during bulk processing. - [How to Seamlessly Switch Video Streams for Many Viewers](https://patentbrief.org/patent/us/6732183/video-and-audio-streaming-for-multiple-users) — This patent describes a computer system that allows an administrator or viewer to smoothly switch between different video or audio sources for many people watching at the same time, without interrupting their viewing experience.