Dynamic Multi-Plane Read Commands for Mixed Workload Parallelism
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems face inefficiencies in processing a combination of sequential and random read commands, leading to reduced performance and increased latency due to the incompatibility of multi-plane read commands with other commands, resulting in idle planes and reduced bus parallelism.
Innovation Solution
A system that dynamically determines whether to output multi-plane or single-plane read commands based on a threshold quantity of planes, and separates multi-plane read commands into individual commands to enhance processing efficiency and parallelism.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multi-plane read commands are used to access multiple planes simultaneously, then throughput is improved, but compatibility with other commands is lost causing planes to become idle
Solution Approach 1:
The patent segments multi-plane read commands into individual single-plane read commands that can be independently processed. This allows each plane to be accessed separately while maintaining compatibility with other command types, resolving the conflict between achieving high throughput through multi-plane operations and maintaining adaptability with various command formats.
Solution Approach 2:
The system dynamically determines whether to issue multi-plane or single-plane read commands based on real-time conditions such as workload characteristics and plane availability. This dynamic adaptation allows the system to optimize for throughput when multi-plane operations are beneficial while falling back to single-plane operations when command compatibility is required, thus resolving the technical contradiction.
2Speed
If multi-plane read commands are issued, then bus parallelism is improved, but command processing flexibility is reduced
Solution Approach 1:
By segmenting multi-plane read commands into separate single-plane commands, the system maintains the ability to process commands flexibly while still achieving bus parallelism through concurrent execution of multiple segmented commands. Each segmented command can be independently scheduled and executed, preserving operational ease while utilizing parallel bus capabilities.
Solution Approach 2:
The system performs preliminary analysis of the read workload to determine the optimal command strategy before execution. This preliminary action allows the system to prepare segmented commands in advance, ensuring both bus parallelism and command processing flexibility are optimized based on the specific workload characteristics.
3Adaptability or versatility
If single-plane read commands are processed individually, then command compatibility is maintained, but processing efficiency is reduced
Solution Approach 1:
The patent merges multiple single-plane read commands into coordinated execution sequences that maintain individual command compatibility while achieving combined processing efficiency. By merging the processing of segmented commands through a unified control mechanism, the system preserves adaptability with various command formats while improving overall processing efficiency through coordinated execution.
4Adaptability or versatility
If the system supports both sequential and random read commands, then workload versatility is improved, but processing complexity increases
Solution Approach 1:
The patent segments the processing of sequential and random read commands into separate handling paths, each optimized for its specific workload type. This segmentation allows the system to maintain versatility in supporting both workload types while reducing processing complexity by applying specialized processing logic to each segment rather than using a single complex processing mechanism for all workloads.
Data Source
AI summary
Methods, systems, and devices for enhancements for multiple data plane read commands are described. In some examples, a memory system may receive a set of multiple read commands and may determine whether a quantity of planes associated with the set of multiple read commands satisfies a threshold. Based on determining that the quantity of planes satisfies a threshold, the memory system may output a multi-plane read command. A memory device may obtain a multi-plane read command and may initiate a first transfer for a first plane based on obtaining the multi-plane read command. The memory device may then generate one or more single-plane commands for one or more planes associated with the multi-plane read command and may initiate a respective data transfer for each of the one or more planes.


