Memory Resource Pooling for Latency Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-performance computing systems face challenges in achieving uniform memory access latencies due to the presence of both volatile and non-volatile DIMMs, leading to resource exhaustion and reduced system performance when accessing slower memory modules.
Innovation Solution
The system manages accesses by allocating resources into pools based on memory access criteria, such as memory address ranges, request types, and node relationships, ensuring that each type of access is directed to appropriate resource pools, preventing resource starvation and allowing other memory accesses to continue unabated.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If non-volatile DIMMs are installed to prevent data loss, then data reliability is improved, but memory access latency increases significantly
Solution Approach 1:
The patent segments memory into different types (volatile and non-volatile DIMMs) with distinct characteristics, allowing the system to utilize both fast volatile memory for performance-critical operations and reliable non-volatile memory for data persistence, thereby resolving the contradiction between speed and reliability
Solution Approach 2:
The patent applies local quality by assigning different memory types to different data requirements - volatile memory for frequently accessed data requiring speed, and non-volatile memory for data requiring persistence, optimizing both latency and reliability for their respective use cases
2Productivity
If resources are shared among all memory accesses, then resource utilization is improved, but resource exhaustion occurs when slow memory accesses occupy resources
Solution Approach 1:
The patent segments shared resources into dedicated pools for different memory types, preventing slow non-volatile memory accesses from exhausting resources needed for fast volatile memory operations, thus maintaining both high resource utilization and availability
Solution Approach 2:
The patent introduces resource pools as intermediary structures that mediate between different memory access types and the underlying hardware resources, allowing the system to manage resource allocation dynamically and prevent exhaustion while maintaining high utilization
3Loss of time
If uniform memory access latency is achieved through identical DIMMs, then access time uniformity is improved, but system adaptability decreases when mixing memory types
Solution Approach 1:
The patent creates a universal memory management system that can handle both volatile and non-volatile DIMMs through a unified interface and resource pool architecture, allowing the system to adapt to different memory configurations while maintaining consistent access patterns and performance characteristics
Data Source
AI summary
A high-performance computing system, method, and storage medium manage accesses to multiple memory modules of a computing node, the modules having different access latencies. The node allocates its resources into pools according to pre-determined memory access criteria. When another computing node requests a memory access, the node determines whether the request satisfies any of the criteria. If so, the associated pool of resources is selected for servicing the request; if not, a default pool is selected. The node then services the request if the pool of resources is sufficient. Otherwise, various error handling processes are performed. Each memory access criterion may relate to a memory address range assigned to a memory module, a type of request, a relationship between the nodes, a configuration of the requesting node, or a combination of these.