C. Pujara, Anurag Modi, G. Sandeep, Shilpa Inamdar, Deepa Kolavil, V. Tholath
{"title":"VC-1 video decoder optimization on ARM Cortex-A8 with NEON","authors":"C. Pujara, Anurag Modi, G. Sandeep, Shilpa Inamdar, Deepa Kolavil, V. Tholath","doi":"10.1109/NCC.2010.5430200","DOIUrl":null,"url":null,"abstract":"Optimization of VC-1 Main profile decoder has been proposed using SIMD engine (NEON) of ARM Cortex-A8 processor. We have exploited data level parallelism to effectively use the SIMD capability of NEON. Ideas to make the algorithms SIMD friendly are also highlighted. With effective use of ARM Cortex- A8 architecture and NEON SIMD engine, the MCPS (Mega Cycles per second) requirement for decoding has reduced considerably as compared to our optimized C implementation. Results are provided for the overall performance gain in decoding and individual performance gain in various modules of VC-1 Main profile decoder.","PeriodicalId":130953,"journal":{"name":"2010 National Conference On Communications (NCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2010-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2010 National Conference On Communications (NCC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/NCC.2010.5430200","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 5
Abstract
Optimization of VC-1 Main profile decoder has been proposed using SIMD engine (NEON) of ARM Cortex-A8 processor. We have exploited data level parallelism to effectively use the SIMD capability of NEON. Ideas to make the algorithms SIMD friendly are also highlighted. With effective use of ARM Cortex- A8 architecture and NEON SIMD engine, the MCPS (Mega Cycles per second) requirement for decoding has reduced considerably as compared to our optimized C implementation. Results are provided for the overall performance gain in decoding and individual performance gain in various modules of VC-1 Main profile decoder.