GPU Encoder Offloading CPU Workload for Content Browsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud-based service systems face limitations in serving multiple clients simultaneously due to high CPU workload and bandwidth occupancy when encoding and streaming raw images, leading to increased power consumption and costs.
Innovation Solution
Leveraging GPU resources by using the graphics card's computation capacity to directly retrieve, encode, and stream raw images from VRAM, reducing the CPU's workload and bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If CPU is used to encode and stream raw images for multiple clients, then service capacity is improved, but CPU workload and bandwidth occupancy increase excessively
Solution Approach 1:
The patent replaces the CPU-based encoding and streaming mechanism with a GPU-based mechanism. The GPU's parallel processing architecture is utilized to encode raw images and stream them to multiple clients simultaneously, substituting the traditional CPU workload with specialized graphics hardware that is better suited for parallel image processing operations.
Solution Approach 2:
The GPU is leveraged to perform multiple functions including encoding raw images, streaming encoded data to multiple clients, and managing video memory operations. This multi-functional approach allows a single hardware component to handle the entire encoding and distribution pipeline, reducing the need for separate processing units and simplifying the system architecture.
2Productivity
If more servers are added to serve more clients simultaneously, then service capacity is improved, but system cost increases
Solution Approach 1:
The patent merges multiple encoding and streaming operations into a single server by utilizing the GPU's capability to handle multiple concurrent encoding tasks. Instead of requiring separate servers for each client or encoding stream, the GPU consolidates these functions, allowing one server to efficiently serve multiple clients simultaneously through parallel processing of image encoding and data streaming.
3Productivity
If CPU encodes and streams video data for multiple clients, then service capacity is improved, but power consumption increases
Solution Approach 1:
The patent substitutes the CPU's general-purpose processing with the GPU's specialized parallel processing architecture optimized for graphics and video operations. This substitution reduces power consumption because the GPU can perform encoding and streaming operations more efficiently per unit of energy, leveraging its dedicated hardware circuits designed specifically for parallel image and video processing tasks.
Data Source
AI summary
A system and a corresponding method of leveraging GPU resources to enhance performance of the interact-able content browsing (ICB) service are to utilize the computation capacity of the GPU of the graphics card to share the computation load originally taken by the CPU. The raw images depicted by the browser application program executed in the server can be directly retrieved from the VRAM of the graphics card by the encoder inside the GPU to be further encoded and streamed before being forwarded directly to distant clients. Thereby, work load and bus bandwidth occupation on the CPU can be greatly reduced, the number of clients that can be served simultaneously by a single server can be increased, and the related power consumption can be substantially reduced.