Multi-Precision Memory Retrieval With Selective Data Combining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The speed at which data can be retrieved from memory becomes a bottleneck, limiting the overall computational throughput as computational speed increases, and existing methods do not efficiently manage data retrieval at multiple levels of precision.
Innovation Solution
A system and method for multi-precision memory retrieval involving a combiner unit that splits data into portions stored across multiple memory interfaces, using multiplexers and buffers to selectively retrieve data at varying precision levels, with optional adjustment values to maintain accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data is retrieved from memory at high precision, then accuracy is improved, but retrieval speed deteriorates
Solution Approach 1:
The patent segments data into multiple precision portions (first precision data and second precision data) stored in separate memory locations. The combiner unit can selectively retrieve only the necessary precision level needed for each computational task, avoiding the overhead of retrieving full-precision data when lower precision suffices, thus improving retrieval speed without sacrificing necessary accuracy.
Solution Approach 2:
The combiner unit dynamically adjusts the precision level of data retrieval based on computational requirements. It can switch between retrieving first precision data, second precision data, or both, depending on the specific computational task, optimizing the balance between speed and accuracy for different operational contexts.
2Productivity
If computational speed increases, then processing capability is improved, but memory retrieval bottleneck worsens
Solution Approach 1:
The system performs preliminary actions by pre-processing data into multiple precision portions and storing them in optimized memory locations before computation. The combiner unit is pre-configured with the ability to quickly assemble different precision levels, reducing the time penalty during actual computation and keeping pace with increased processing capability.
Solution Approach 2:
The patent changes the parameter of data precision by storing and retrieving data at different precision levels. This allows the system to match memory retrieval bandwidth with computational requirements, preventing the memory bottleneck from limiting overall productivity when computational speed increases.
3Adaptability or versatility
If data is partitioned into multiple portions, then retrieval flexibility is improved, but system complexity worsens
Solution Approach 1:
The combiner unit is designed as a universal component that can handle multiple precision retrieval scenarios using the same basic architecture. It universally manages first precision data, second precision data, and their combinations through a standardized interface, reducing the effective complexity despite the partitioned data structure.
Data Source
AI summary
A system for retrieving data at multiple precision levels includes a processor, a memory device for storing a first portion of data in association with a first interface and a second portion of data in association with a second interface, and a combiner unit. The combiner unit is communicably coupled to the memory device and the processor and includes a first buffer coupled to the first interface for receiving the first portion of data; and a second buffer coupled to the second interface for receiving the second portion of data. The combiner unit further includes selectors coupled to the first and second buffers. The one or more selector units select for output the first portion of data based on the combiner unit being in a first state and the first portion of data and the second portion of data based on the combiner unit being in a second state.