Media Hub Subsystem Local Cache Bandwidth Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current computing systems face challenges in reducing DRAM bandwidth consumption and power management across System-on-Chip (SoC) due to ineffective connection of media hardware accelerators, leading to increased platform costs, power, and thermal stress during media processing.

Innovation Solution

Integration of media accelerators into a media hub subsystem with a local cache memory for inter-media accelerator communication, enabling stripe caching of reference frames for video encoding, decoding, and post-processing, which reduces bandwidth and power consumption by minimizing direct memory access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple media hardware accelerators access memory directly, then media processing capability is improved, but DRAM bandwidth consumption increases significantly

Engineering Contradiction:
Improvemedia processing capabilityVSAvoidDRAM bandwidth consumption
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent merges multiple media hardware accelerators (video decoder, video encoder, image signal processor) into a unified media hub subsystem that shares a common local cache memory. This consolidation allows the accelerators to communicate with each other through the local cache rather than directly accessing DRAM, thereby maintaining high media processing capability while significantly reducing DRAM bandwidth consumption.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The local cache memory acts as an intermediary between the media hardware accelerators and the DRAM. By buffering data in the local cache, the mediators enable the accelerators to exchange data efficiently without requiring frequent DRAM accesses, thus reducing the bandwidth burden on DRAM while preserving processing throughput.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If media accelerators are not effectively connected, then system complexity is reduced, but power management across SoC deteriorates

Engineering Contradiction:
Improvesystem complexityVSAvoidpower management
Core Design Contradiction:
Device complexityVSUse of energy by stationary object

Solution Approach 1:

The patent combines multiple media accelerators and a shared local cache into an integrated media hub subsystem. This merging creates an effective connection among the accelerators, enabling coordinated operation and improved power management across the SoC, as the subsystem can be managed as a unified entity with shared resources.

Inventive Principle:
Principle #5Merging (Combining)

3Manufacturing precision

If DRAM bandwidth is increased to support higher resolution media, then media quality is improved, but platform cost and thermal stress increase

Engineering Contradiction:
Improvemedia qualityVSAvoidthermal stress
Core Design Contradiction:
Manufacturing precisionVSObject-generated harmful factors

Solution Approach 1:

The local cache memory serves as an intermediary buffer that allows high-resolution media data to be processed efficiently without requiring proportionally high DRAM bandwidth. By caching frequently accessed data locally, the system can maintain high media quality while reducing the thermal load associated with high-speed DRAM access.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3284263B1Media hub device and cache
Publication Date: 2021.06.30 INTEL CORP
  • EP3284263B1 patent drawingFigure 1
  • EP3284263B1 patent drawingFigure 2
  • EP3284263B1 patent drawingFigure 3

AI summary

The present disclosure describes techniques related to media caching. A media hub device may include a media hub device configured to execute an operation on a current frame of media having a frame period. The media hub device may include a cache configured to provide, to a media accelerator of the media hub device, data associated with the frame period of the current frame.