Neural Processing Unit Delegation Control via Privilege Levels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing neural processing units (NPUs) require a direct hardware interface for each program, leading to increased hardware complexity and inefficiency when multiple programs need to access resources, as existing access management schemes can be slow and cumbersome.
Innovation Solution
A neural processing unit with a delegation mechanism that allocates transactions to delegation pages based on program and page privilege levels, allowing higher privilege programs to arbitrate access to resources, reducing the need for direct interfaces and enabling dynamic access control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a direct hardware interface is provided for each program, then access to resources is enabled, but hardware complexity increases
Solution Approach 1:
The patent implements a universal delegation page structure that can serve multiple programs with different privilege levels. Instead of creating dedicated hardware interfaces for each program, a single delegation page framework handles access control for all programs, making the system multi-functional and reducing overall hardware complexity.
Solution Approach 2:
The delegation page acts as an intermediary layer between programs and resources. Rather than direct program-to-resource interfaces, the delegation page mediates access by receiving transactions from programs, determining appropriate resource access based on privilege levels, and managing the actual resource allocation, thereby reducing the need for multiple direct interfaces.
2Device complexity
If access management is centralized for all programs, then hardware complexity is reduced, but access speed decreases
Solution Approach 1:
The patent segments the centralized access management into multiple delegation pages, each handling specific privilege levels and resource types. This segmentation allows parallel processing of transactions from different programs at appropriate privilege levels, maintaining fast access speeds while keeping the overall system simplified through the delegation framework.
Data Source
AI summary
A neural processing unit comprises an input module for receiving a transaction from at least one program, each program has an associated program privilege level; and a plurality of delegation pages, each delegation page comprising a delegation management unit associated with a page privilege level. The neural processing unit also comprises at least one resource arranged to be accessed by at least one of the delegation pages; and a processing module arranged to process the transaction. Processing the transactions comprises allocating each transaction to a delegation page based on the program privilege level and page privilege level. The program is arranged to instruct the delegation management unit of a first delegation page, having a first-page privilege level to delegate access to the at least one resource to a second delegation page having a second-page privilege level, and wherein the first-page privilege level is higher than the second-page privilege level.