GPU Firmware Method for CPU Throughput

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional graphics processing unit (GPU) software methods require an exchange of information with the central processing unit (CPU), leading to reduced throughput due to GPU idling and CPU interrupts, which decreases overall processing efficiency.

Innovation Solution

Implementing a firmware method executed by a processor within the GPU that performs software functions without exchanging information with the CPU, minimizing GPU idling and CPU interrupts, thereby enhancing processing throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If software methods are executed in conventional GPU by exchanging information with CPU, then GPU state can be examined and updated, but GPU processing is idled and CPU throughput is reduced

Engineering Contradiction:

ImproveGPU state management accuracyVSAvoidGPU processing throughput

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent extracts the software method execution function from the CPU and relocates it to a processor within the GPU. This allows the GPU to examine and update its own state independently without requiring CPU intervention, thereby eliminating the idle periods caused by CPU-GPU information exchange while maintaining accurate state management

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The GPU is equipped with its own processor that can autonomously execute software methods and manage its internal state. This self-service capability allows the GPU to perform configuration updates and state examinations without external CPU assistance, eliminating the bottleneck of CPU-GPU communication and maintaining continuous processing throughput

Inventive Principle:

Principle #25Self-service

2Ease of operation

If CPU interrupts are generated to execute software methods, then GPU configuration can be updated, but CPU throughput is significantly reduced

Engineering Contradiction:

ImproveGPU configuration flexibilityVSAvoidCPU throughput

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent extracts the software method execution capability from the CPU and places it within the GPU's own processor. This eliminates the need for CPU interrupts to configure the GPU, as the GPU can now autonomously execute software methods and update its configuration without interrupting CPU operations

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces an intermediate processor within the GPU that acts as a mediator between software methods and the GPU hardware. This intermediate processor handles software method execution and GPU configuration updates locally, eliminating the need for CPU intervention and preserving CPU throughput while maintaining configuration flexibility

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS9734545B2Software methods in a GPU

Publication Date: 2017.08.15 NVIDIA CORP

AI summary

One embodiment of the present invention sets forth a technique for executing a software method within a graphics processing unit (GPU) that minimizes the number of clock cycles during which the graphics engine is idled. The function of the software method is performed by a firmware method that is executed by a processor within the GPU. The firmware method is executed to access and optionally update the state stored in the GPU. Unlike execution of a conventional software method, execution of the firmware method does not require an exchange of information between a CPU and the GPU. Therefore, the CPU is not interrupted and throughput of the CPU is not reduced.