Many application domains of computer science, including: astrophysics, molecular dynamics, quantum chemistry, image processing, financial simulation, computer tomography, structural biology as well as many other require a lot of computing power. Fortunately, in most cases it is possible to parallelize parts of algorithms used to solve the problems in the aforementioned domains. However, even if parallelized, many of these algorithms require powerful and expensive hardware capable of running many parallel communicating and cooperating threads.
Recently however, the scientific community has realized, that Graphics Processing Units (GPUs) of modern graphics cards have similar capabilities. GPUs turned out to be very powerful (and cheap) computing hardware that may be used in all kinds of computation (not related to computer graphics). Since then, it has been shown many times over, that by utilizing the GPUs we are able to improve the computing efficiency in many computer science application domains and achieve speedups of 20-100X. However, while other fields of computer science already strongly benefit from utilizing GPUs, there is not yet much research related to using them in databases.
The concept of using GPUs in databases is relatively young and has not yet received enough attention from the database community. It is therefore necessary to popularize GPUs amongst the database scientific community and the best way to do that is to organize a workshop. Aside from popularizing the GPUs, our workshop will also provide a forum for discussion amongst the people interested in the subject. We think that organizing a workshop devoted to a relatively young field of computer science, which has the potential of allowing to achieve high speedups in many applications domains of databases is a very attractive idea.
February 7th, 2014
GID 2014 website started.