• The Journal of the Korea Contents Association
• /
• v.14 no.3
• /
• pp.22-32
• /
• 2014

Recently, GPGPU has been widely used for general-purpose processing as well as graphics processing by providing optimized hardware for parallel processing. Memory system has big effects on the performance of parallel processing units such as GPU. In the GPU, hierarchical memory architecture is implemented for high memory bandwidth. Moreover, both memory address coalescing and memory request merging techniques are widely used. This paper analyzes the GPU performance according to various memory organizations. According to our simulation results, GPU performance improves by 15.5%, 21.5%, 25.5%, 30.9% as adding 8KB L1, 16KB L1, 32KB L1, 64KB L1 cache, respectively, compared to case without L1 cache. However, experimental results show that some benchmarks decrease performance since memory transaction increases due to data dependency. Moreover, average memory access latency is increased as the depth of hierarchical cache level increases when cache miss occurs significantly.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.8
• /
• pp.9-17
• /
• 2019

In this paper, we propose a block mapping technique applicable to NAND flash memory. In order to use the NAND flash memory with the operating system and the file system developed on the basis of the hard disk which is mainly used in the general PC field, it is necessary to use the system software known as the FTL (Flash Translation Layer). FTL overcomes the disadvantage of not being able to overwrite data by using the address mapping table and solves the additional features caused by the physical structure of NAND flash memory. In this paper, we propose a new mapping method based on the block mapping method for efficient use of the NAND flash memory. In the case of the proposed technique, the data modification operation is processed by using a blank page in the existing block without using an additional block for the data modification operation, thereby minimizing the block unit deletion operation in the merging operation. Also, the frequency of occurrence of the sequential write request and random write request Accordingly, by optimally adjusting the ratio of pages for recording data in a block and pages for recording data requested for modification, it is possible to optimize sequential writing and random writing by maximizing the utilization of pages in a block.