基于GPU的Viterbi算法在线语音解码优化策略

2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA) Pub Date : 2018-08-01 DOI:10.1109/ICAICTA.2018.8541343

Alfonsus Raditya Arsadjaja, Achmad Imam Kistijantoro

{"title":"基于GPU的Viterbi算法在线语音解码优化策略","authors":"Alfonsus Raditya Arsadjaja, Achmad Imam Kistijantoro","doi":"10.1109/ICAICTA.2018.8541343","DOIUrl":null,"url":null,"abstract":"Automatic Speech Recognition (ASR) has been popular recently. But the current algorithm for speech recognition is slow and needed the way to recognize faster. One way to achieve it is with GPU, which provides parallel computation; but ASR is hard to parallelize directly.This paper describes how to build parallel ASR system, which requires several steps. First, we must convert the data structure to make it compatible with GPU, then we have to make several kernels that equivalent to the serial algorithm in CPU.We will describe several optimization strategies for make ASR run much faster after we got the correct GPU program. Those strategies are based on profiling result and analysis of the GPU program execution flow.Best implementation that we had have a speedup around 5.59-6.18 times from the serial CPU implementation.","PeriodicalId":184882,"journal":{"name":"2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA)","volume":"148 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Online Speech Decoding Optimization Strategy with Viterbi Algorithm on GPU\",\"authors\":\"Alfonsus Raditya Arsadjaja, Achmad Imam Kistijantoro\",\"doi\":\"10.1109/ICAICTA.2018.8541343\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Automatic Speech Recognition (ASR) has been popular recently. But the current algorithm for speech recognition is slow and needed the way to recognize faster. One way to achieve it is with GPU, which provides parallel computation; but ASR is hard to parallelize directly.This paper describes how to build parallel ASR system, which requires several steps. First, we must convert the data structure to make it compatible with GPU, then we have to make several kernels that equivalent to the serial algorithm in CPU.We will describe several optimization strategies for make ASR run much faster after we got the correct GPU program. Those strategies are based on profiling result and analysis of the GPU program execution flow.Best implementation that we had have a speedup around 5.59-6.18 times from the serial CPU implementation.\",\"PeriodicalId\":184882,\"journal\":{\"name\":\"2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA)\",\"volume\":\"148 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2018-08-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICAICTA.2018.8541343\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICAICTA.2018.8541343","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 1

摘要

自动语音识别技术(ASR)近年来得到了广泛的应用。但目前的语音识别算法速度较慢，需要更快的识别方法。实现它的一种方法是使用GPU，它提供并行计算;但是ASR很难直接并行化。本文介绍了如何构建并行ASR系统，这需要几个步骤。首先，我们必须转换数据结构使其与GPU兼容，然后我们必须在CPU中制作几个相当于串行算法的内核。在得到正确的GPU程序后，我们将描述几种优化策略，使ASR运行得更快。这些策略是基于分析结果和对GPU程序执行流程的分析。我们所拥有的最佳实现比串行CPU实现的速度提高了5.59-6.18倍。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Online Speech Decoding Optimization Strategy with Viterbi Algorithm on GPU

Automatic Speech Recognition (ASR) has been popular recently. But the current algorithm for speech recognition is slow and needed the way to recognize faster. One way to achieve it is with GPU, which provides parallel computation; but ASR is hard to parallelize directly.This paper describes how to build parallel ASR system, which requires several steps. First, we must convert the data structure to make it compatible with GPU, then we have to make several kernels that equivalent to the serial algorithm in CPU.We will describe several optimization strategies for make ASR run much faster after we got the correct GPU program. Those strategies are based on profiling result and analysis of the GPU program execution flow.Best implementation that we had have a speedup around 5.59-6.18 times from the serial CPU implementation.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2018 5th International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA)

自引率

0.00%

发文量