Disaggregated Accelerator Resource Allocation Without Workload Interruptions

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing accelerator devices face inefficiencies in dynamically reallocating shared resources like high-bandwidth memory, leading to delays and suboptimal performance due to static allocation methods that require workload interruption for reconfiguration.

Innovation Solution

Implementing a system with an orchestrator server and dynamic resource allocation logic within accelerators to monitor and adjust resource utilization thresholds in real-time, allowing for dynamic allocation of shared resources among logic portions without interrupting workload execution.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If static resource allocation is used among logic portions, then resource allocation is simple and stable, but resource utilization efficiency deteriorates when workload needs change

Engineering Contradiction:

Improveresource utilization efficiencyVSAvoidresource allocation complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements dynamic resource allocation by allowing logic portions to access shared high bandwidth memory resources on-demand without static pre-allocation. The memory interface dynamically routes memory access requests from multiple logic portions to the shared memory pool, enabling resource allocation to adapt automatically to changing workload requirements while maintaining system stability.

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If reconfiguration is performed to reallocate shared resources, then resource allocation adapts to workload needs, but execution time increases due to stopping and resuming workloads

Engineering Contradiction:

Improveresource allocation adaptabilityVSAvoidworkload execution time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent establishes a shared high bandwidth memory pool in advance that is accessible by multiple logic portions simultaneously. This preliminary setup eliminates the need for reconfiguration during workload execution, as the memory resources are pre-configured to be shared among all logic portions through a common memory interface that handles dynamic access requests.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent enables continuous workload execution by implementing a memory interface that dynamically manages access to the shared memory pool without interrupting logic portion operations. The interface continuously processes memory access requests from multiple logic portions concurrently, ensuring uninterrupted workload execution while adapting memory allocation to changing needs in real-time.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS11336547B2Technologies for dynamically managing resources in disaggregated accelerators

Publication Date: 2022.05.17 INTEL CORP

AI summary

Technologies for dynamically managing resources in disaggregated accelerators include an accelerator. The accelerator includes acceleration circuitry with multiple logic portions, each capable of executing a different workload. Additionally, the accelerator includes communication circuitry to receive a workload to be executed by a logic portion of the accelerator and a dynamic resource allocation logic unit to identify a resource utilization threshold associated with one or more shared resources of the accelerator to be used by a logic portion in the execution of the workload, limit, as a function of the resource utilization threshold, the utilization of the one or more shared resources by the logic portion as the logic portion executes the workload, and subsequently adjust the resource utilization threshold as the workload is executed. Other embodiments are also described and claimed.