Font Size: a A A

Performance Analysis And Optimization Of EMMC-Based Storage

Posted on:2013-04-14Degree:MasterType:Thesis
Country:ChinaCandidate:W W HuFull Text:PDF
GTID:2248330371967706Subject:Computer technology
Abstract/Summary:PDF Full Text Request
From 2009,The mobile phone markets were filled with the wave of smart phones, and the use of eMMC(embedded Multimedia Card) is more and more widely too. However, due to the characteristics of eMMC device, eMMC storage system are faced with several problems, such as devices from different manufactures vary widely on read and write performance; The random read and random write performance is unstable and poor, especially for small random data write performance; Many devices’life are unstable. Therefore, we urgently need to optimize the performance of embedded memory systems, in order to improve the smart phone storage system performance. At present, although the optimization of the storage system performance does not have a uniform standard, but their goal is clear, meeting the user’s read and write requests as quickly as possible.Firstly, this paper introduces the implementation of block device driver and the principle and workflow of the device driver. Secondly, analyzing the characteristics of the device eMMC, and the main features of eMMC driver. Then this paper introduce three new driver features based on original driver to improve the performance of eMMC and extend the life of eMMC. The three new features are TRIM^ HPI(high priority interrupt) and BO(Background Operation). This paper also analyze the I/O scheduling algorithm of embedded Linux system, and based on I/O scheduler introduce a software program, which can improve the small block random write performance of eMMC device greatly. Finally, the author choose the specific testing tools and did the actual reading and writing tests after optimization, and achieve the expected results.
Keywords/Search Tags:Storage systems, embedded system, embedded Multimedia Card, device driver, optimization
PDF Full Text Request
Related items