Macro Cell Video Compression via Cross-Linked GPU Raster Devices

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current video compression technologies for high contrast and natural imagery on desktop computers are costly, power-intensive, and suffer from bandwidth limitations, leading to high latency and poor frame rate performance, which is unsustainable for low-cost, low-power applications such as remote desktop connectivity and aircraft black box recording.

Innovation Solution

A macro cell video compression system utilizing two or more GPUs with cross-linked raster devices for frame locking and pixel locking, generating macro cells by stalling the pixel clock and using Start Address registers to create side-by-side sub-windows, producing checksums and descriptive data for improved compression, and employing full frame and column macro cell scans to avoid tearing effects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional video compression hardware is used, then compression capability is achieved, but cost and power consumption increase significantly

Engineering Contradiction:

Improvevideo compression capabilityVSAvoidpower consumption

Core Design Contradiction:

ReliabilityVSUse of energy by stationary object

Solution Approach 1:

The patent reuses existing GPU raster scan circuits for their original display function while simultaneously enabling video compression functionality. The same hardware circuits perform both raster scanning for display and macro-cell generation for compression, eliminating the need for dedicated compression hardware and reducing power consumption.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The GPU's existing raster scan circuits serve themselves by performing compression operations during normal display operations. The system uses its own internal resources (raster circuits, memory bandwidth) to accomplish compression without requiring external dedicated compression hardware, thereby reducing overall system power consumption.

Inventive Principle:

Principle #25Self-service

2Ease of manufacture

If software-based video compression is used, then hardware cost is reduced, but frame rate performance and latency worsen

Engineering Contradiction:

Improvehardware costVSAvoidframe rate performance

Core Design Contradiction:

Ease of manufactureVSProductivity

Solution Approach 1:

The patent replaces software-based compression algorithms with hardware-based circuit implementations within the GPU. The raster scan circuits and macro-cell generation logic are implemented as hardware circuits that operate in parallel, providing hardware-speed performance without requiring expensive dedicated compression hardware.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent combines video compression functionality with the existing GPU raster scan hardware. By merging compression operations with the display pipeline, the system achieves hardware-accelerated compression performance using existing high-speed graphics circuits rather than relying on slower software processing.

Inventive Principle:

Principle #5Merging (Combining)

3Productivity

If dedicated video compression hardware is used, then compression performance is improved, but device complexity and cost increase

Engineering Contradiction:

Improvecompression performanceVSAvoidhardware complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent makes the GPU's raster scan circuits multi-functional by enabling them to perform both display operations and video compression operations. The same hardware blocks generate raster data for display while simultaneously generating macro-cells for compression, reducing overall device complexity without sacrificing compression performance.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent uses the GPU's existing memory interface and data buses as intermediaries to transfer data between the raster scan circuits and the compression logic. This approach avoids creating separate dedicated data paths and reduces hardware complexity by leveraging existing communication infrastructure within the GPU.

Inventive Principle:

Principle #24Intermediary (Mediator)

4Loss of time

If high-speed compression is implemented, then latency is reduced, but power consumption increases

Engineering Contradiction:

ImprovelatencyVSAvoidpower consumption

Core Design Contradiction:

Loss of timeVSUse of energy by stationary object

Solution Approach 1:

The patent performs compression operations continuously during the raster scan process without requiring separate processing stages. As the raster circuits continuously scan and generate display data, the same circuits continuously generate macro-cells for compression, achieving low latency through continuous operation while maintaining power efficiency by using existing circuitry throughout the process.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS10499072B2Macro cell display compression multi-head raster GPU

Publication Date: 2019.12.03 MIMAX INC

AI summary

A macro cell video compression system and related method, the system including a first raster device and a second raster device that is cross-linked to the first raster device to move macro cells between the linked first and second raster devices, each of the first and second raster devices comprising a video raster circuit having a raster phase locking circuit to perform frame locking and pixel locking between the first and second raster devices and to generate macro cells by one of the first and second raster devices, each of the first and second raster devices has two or more raster heads that generate the macro cells by pixel clock stalling.