Adaptive Sequence Alignment via Hybrid CPU-GPU Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Sequence alignment applications in bioinformatics face significant computation time challenges due to the growing size of sequence data, with existing methods being inefficient in processing large databases, especially when dealing with sequences of varying lengths.
Innovation Solution
A hybrid GPU/CPU system is employed, where long sequences are processed on the GPU and short sequences on the CPU, optimizing the distribution of database sequences between processors to enhance processing efficiency and reduce computation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all sequences are processed on a single processor type, then device complexity is reduced, but productivity decreases due to inefficient processing of sequences with varying lengths
Solution Approach 1:
The patent segments the sequence processing workload by length thresholds, directing long sequences to GPU processing and short sequences to CPU processing. This segmentation allows each processor type to operate in its optimal performance range, thereby improving overall productivity without requiring a complete system redesign.
Solution Approach 2:
The system dynamically routes sequences to different processors based on their length characteristics. The adaptive nature of the system allows it to adjust processing distribution in real-time based on sequence properties, optimizing performance while maintaining manageable system complexity through flexible rather than rigid architecture.
2Loss of time
If sequences of varying lengths are processed uniformly, then ease of operation is maintained, but loss of time increases due to suboptimal processing efficiency
Solution Approach 1:
The patent applies different processing qualities to different sequence segments based on their characteristics. Long sequences receive GPU-based optimized processing while short sequences use CPU-based processing, allowing each segment to receive the appropriate level of computational resources and optimization, thereby reducing overall computation time without complicating the user interface.
Solution Approach 2:
The system automatically classifies and routes sequences to appropriate processors based on their length, eliminating the need for manual intervention or complex user configuration. This self-service approach reduces computation time through automatic optimization while maintaining ease of operation for the end user.
3Productivity
If GPU processing is used for all sequences, then productivity improves for long sequences, but device complexity increases due to need for handling varying sequence lengths
Solution Approach 1:
The patent segments the workload by sequence length, assigning long sequences to GPU processing for high throughput and short sequences to CPU processing for efficient management. This segmentation resolves the complexity of handling varying sequence lengths on a single processor type while maintaining high productivity for the most computationally intensive tasks.
Solution Approach 2:
The system changes the processing parameter (processor type) based on sequence length parameters. By adjusting which processor handles which sequences based on their characteristics, the system optimizes productivity for long sequences while managing the complexity of varying workloads through parameter-based routing rather than complex adaptive algorithms.
4Productivity
If CPU processing is used for all sequences, then device complexity is minimized, but productivity decreases significantly for long sequences
Solution Approach 1:
The patent segments sequences by length and distributes them to appropriate processors, allowing long sequences to benefit from GPU acceleration for improved productivity while keeping CPU processing for short sequences. This segmentation approach increases overall processing speed without requiring complete system complexity, as only the necessary portion of the workload is handled by the more complex GPU path.
Data Source
AI summary
Disclosed are various embodiments for adaptive processing for sequence alignment. In one embodiment, among others, a method includes obtaining a query sequence and a plurality of database sequences. A first portion of the plurality of database sequences is distributed to a central processing unit (CPU) and a second portion of the plurality of database sequences is distributed to a graphical processing unit (GPU) based upon a predetermined splitting ratio associated with the plurality of database sequences, where the database sequences of the first portion are shorter than the database sequences of the second portion. A first alignment score for the query sequence is determined with the CPU based upon the first portion of the plurality of database sequences and a second alignment score for the query sequence is determined with the GPU based upon the second portion of the plurality of database sequences.