2013, Vol.18, No.11

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

pISSN : 1598-849X / eISSN : 2383-9945
https://journal.kci.go.kr/jksci

In multi-core processors, Last Level Cache(LLC) can reduce the speed gap between the memory and the core. For this reason, LLC has big impact on the performance of processors. LLC is composed of shared cache and private cache. In computer architecture community, most researchers have mainly focused on the management techniques for shared cache, while management techniques for private cache have not been widely researched. In conventional private LLC, memory is statically assigned to each core, resulting in serious performance degradation when the workloads are not fairly distributed. To overcome this problem, this paper proposes the replacement policy for managing private cache of LLC efficiently. As proposed core-aware cache replacement policy can reconfigure LLC dynamically, hit rate of LLC is increases drastically. Moreover, proposed policy uses 2-bit saturating counters to improve the performance. According to our simulation results, the proposed method can improve hit rates by 9.23% and reduce the access time by 12.85% compared to the conventional method.

The goal of this paper is to provide performance improvement and stability for satellite data correction of some distortions due to cloud or radiance through distributed processing on cluster. To do this, we proposed and implemented SGE(Sun Grid Engine) based distributed processing methods using local storages and a status table. In the verification, the experiment result revealed that the proposed system on seven nodes improved the processing speed by 138.81% as compare to the existing system and provided good stability as well. This result showed that the proposed distributed processing work is more appropriate to process CPU bound jobs than I/O bound jobs. We expect that the proposed system will give scientists improved analysis performance in various fields and near-real time analysis services.

In this paper, we propose pre-arrangement based task scheduling scheme to reduce MapReduce job processing time. If a task and data to be processed do not locate in same node, the data should be transmitted to node where the task is allocated on. In that case, a job processing time increases owing to data transmission time. To avoid that case, we schedule tasks into two steps. In the first step, tasks are sorted in the order of high data locality. In the second step, tasks are exchanged to improve their data localities based on a location information of data. In performance evaluation,we compare the proposed method based Hadoop with a default Hadoop on a small Hadoop cluster in term of the job processing time and the number of tasks sorted to node without data to be processed by them. The result shows that the proposed method lowers job processing time by around 18%. Also, we confirm that the number of tasks allocated to node without data to be processed by them decreases by around 25%.