Hardware Accelerator Cache Mapping for Process Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional hardware accelerators in multi-core systems lack the ability to track process migration, leading to inefficiencies in writing operation results to the correct cache memory, resulting in performance penalties due to cache coherency operations.
Innovation Solution
Incorporating a mapping table within the hardware accelerator that maps processor identifiers to cache memory identifiers, allowing the accelerator to update the destination cache for migrated software processes, ensuring results are written to the appropriate cache for faster access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional hardware accelerators write results to the original cache memory location without tracking process migration, then the hardware accelerator structure remains simple, but the system performance deteriorates due to cache coherency operations
Solution Approach 1:
The patent introduces a process identifier register and a mapping table as intermediary components between the hardware accelerator and cache memory systems. The process identifier register captures the PID from the calling processor, and the mapping table translates this PID to the current cache memory location, enabling the accelerator to adapt to process migration without increasing overall system complexity
Solution Approach 2:
The hardware accelerator performs preliminary actions by capturing the process identifier at the time of the operation call and proactively looking up the current cache memory location before writing results. This preliminary identification and mapping process ensures that results are always written to the correct cache location, preventing performance degradation from cache coherency operations
2Ease of operation
If hardware accelerators use cache coherency operations to access operation results from other caches, then cache memory access remains simple, but the speed and efficiency deteriorate
Solution Approach 1:
The patent implements a feedback mechanism where the hardware accelerator continuously monitors the process identifier and uses the mapping table to determine the current cache memory location of the migrated process. This feedback loop ensures that the accelerator always writes to the correct cache location, eliminating the need for cache coherency operations and maintaining high access speeds
3Speed
If the hardware accelerator tracks process migration and updates cache destination, then the access speed to operation results improves, but the device complexity increases
Solution Approach 1:
The patent segments the cache tracking function into distinct modular components: a process identifier register for capturing PIDs, a mapping table for PID-to-cache location translation, and control logic for updating the destination cache. This segmentation allows each component to perform its specific function independently, improving access speed while managing complexity through modular design
Data Source
AI summary
A mechanism is provided by which a hardware accelerator detects migration of a software process among processors and uses this information to write operation results to an appropriate cache memory for faster access by the current processor. This mechanism is provided, in part, by incorporation within the hardware accelerator of a mapping table having entries including a cache memory identifier associated with a processor identifier. The hardware accelerator further includes circuitry configured to receive a processor identifier from a calling processor, and to perform a look-up in the mapping table to determine the cache memory identifier associated with the processor identifier. The hardware accelerator uses the associated cache memory identifier to write results of called operations to the cache memory associated with the calling processor, thereby accelerating subsequent operations by the calling processor that rely upon the hardware accelerator results.