Our research topics are including but not limited to the following.
Our goal is to develop efficient computing methods that utilize the latest hardware technologies.
In recent years, the access gap between main memory and CPU has become larger. This problem is known as the “memory wall problem”. As a CPU has multicores, the efficiency of concurrency control becomes more important for programs that share a large amount of data. On the other hand, as a graphics processor (GPU) becomes programmable, it is now possible to compute a large amount of data in parallel with its many simplified processors. However, since a GPU is designed for numerical computations, it is difficult to use them for many non-numerical operations, such as text processing.
In this research, we will develop methods to efficiently perform large-scale computations by utilizing multicore CPUs, GPUs, and new memory technologies, from the viewpoints of software, algorithms, and hardware, and through their coordination.
Data Access Methods for Big Data
We are studying efficient data access technologies to utilize big data in applications.
In recent years, the use of diverse and huge data, or big data, has attracted much attention. Such data is stored in cloud storage with many computers. Although cloud storage can handle a large amount of data, careful design is required for efficient data access from applications and services in advance. Moreover changes in access to the data often result in inefficient performance.
In this research, we will develop technologies to provide efficient data access even when any types of new applications and services join to use cloud storage.
Other Data-oriented Technologies
We aim to solve many issues in core technologies, such as database systems, and applications that manage and utilize a large amount of data. Past research topics can be found in our publication list.