Many time series algorithms reduce the computation cost by pruning unpromising candidates with lower-bound distance functions. In this paper, we focus on an orthogonal research direction that further boosts the performance by unlocking the potentials of modern commodity CPUs. First, we conduct a performance profiling on existing algorithms to understand where does time go. Second, we design vectorized implementations for lower-bound and distance functions that can enjoy characteristics (e.g., data parallelism, caching, branch prediction) provided by CPU. Third, our vectorized methods are general and applicable to many time series problems such as subsequence search, motif discovery and kNN classification. Our experimental study on real datasets shows that our proposal can achieve up to 6 times of speedup.
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||7th International Workshop on Accelerating Data Analysis and Data Management Systems Using Modern Processor and Storage Architectures, ADMS 2016 and 4th International Workshop on In-Memory Data Management and Analytics, IMDM 2016|
|Period||1/09/16 → 1/09/16|
- Theoretical Computer Science
- Computer Science(all)