Super Workload Dispatch for Single Context Processor Utilization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Single context multicore processors are underutilized due to limited parallelism in certain applications like video encoding, leading to inefficiency and performance scalability issues, as they cannot fully utilize available execution units despite having more units than required for the task.

Innovation Solution

Creating a super workload by grouping independent workloads together and dispatching them as a single entity to a single context hardware system, allowing for efficient distribution across multiple execution units, thereby achieving multi-context processing without the need for hardware modifications.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single context multicore processor is used, then hardware complexity is reduced, but execution unit utilization is insufficient

Engineering Contradiction:

Improvehardware complexityVSAvoidexecution unit utilization

Core Design Contradiction:

Device complexityVSProductivity

Solution Approach 1:

The patent merges multiple independent workloads into a single assembled workload that is dispatched to a single context hardware processor. This combining approach allows the processor to handle multiple workloads sequentially within a single context, achieving high execution unit utilization without requiring complex multi-context hardware architecture.

Inventive Principle:

Principle #5Merging (Combining)

2Productivity

If more execution units are added to the system, then performance scalability is improved, but hardware cost and complexity increase

Engineering Contradiction:

Improveperformance scalabilityVSAvoidhardware cost

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent ensures continuous utilization of execution units by assembling multiple independent workloads into a single workload that keeps all execution units busy. This approach achieves performance scalability by maximizing the utilization of existing execution units rather than adding more hardware, thereby avoiding increased hardware cost and complexity.

Inventive Principle:

Principle #20Continuity of useful action

3Productivity

If algorithm changes are made to increase parallelism, then execution unit utilization is improved, but result quality decreases

Engineering Contradiction:

Improveexecution unit utilizationVSAvoidresult quality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent segments the assembled workload into multiple independent workloads that can be processed sequentially. Each workload maintains its original algorithm integrity and spatial dependencies, ensuring result quality is not compromised while achieving high execution unit utilization through the assembly and dispatch mechanism.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11150943B2Enabling a single context hardware system to operate as a multi-context system

Publication Date: 2021.10.19 INTEL CORP

AI summary

Independent workloads may be grouped together into a single super workload. This super workload is dispatched to a single context hardware system that does not run an operating system. This effectively creates a multi-context system out of a single context hardware processor.