How to access a database parallely using CUDA ?

I am currently working on a project where i need to access the database parallely using NVIDIA CUDA.
I require each thread to fetch each entry within the database parallely.
I am not able to figure out how this should be done exactly.
Can anyone help me or guide me on this issue ?

[1] What sort of database?
[2] What have you tried so far? Where did you run into problems?
[3] What literature did you consult?
[4] Have you discussed possible approaches with your advisor or mentor?

[1]It is supposed to be SQL database.
[2]I have not yet tried anything because i could not find anything related to connecting an SQL database with CUDA.
[3]I am trying to achieve this:
http://link.springer.com/article/10.1007/s11227-011-0672-7
[4]Yes

When I search for “database sql gpu cuda” at Google Scholar (https://scholar.google.com/scholar?q=database+sql+gpu+cuda&btnG=&hl=en&as_sdt=0%2C5) it find dozens of papers that look relevant.

It is not clear what literature you have consulted other than the one paper you cited, but I would recommend performing a thorough literature search first, and working through all relevant papers you can get our hands on. This will give you a feel for the state of the art as well as the recent history of the field.

You would also want to familiarize yourself with CUDA through practical programming exercises, if you aren’t fluent in CUDA yet. In conjunction with discussions with your advisor, that should allow you to devise a strategy for your project. Expect lots of hard work ahead, such is the nature of projects in as of yet unfamiliar application areas.

I am not familier with CUDA APIs which needs to be used.
I am lacking the skill of CUDA programming in this context.
So can you suggest me any book or any other source which would help me get familier to CUDA programming ?

If you can program in C, the canonical source of documentation is a very good one.

Thank You for the link.
I really appriciate it.
One of the major problems which i am facing is that i do not have an NVIDIA card.Can u suggest me something which would help me do CUDA programming on the same machine ?

Buy a CUDA capable card, or use OpenCL instead (the latter not necessarily recommended as CUDA seems to have a better community, like this site).

Does your supervisor have experience with CUDA? I’d be interested in a reply to njuffa’s question [4] that goes a bit beyond a simple yes/no.

yes i have discussed it with my mentor.
He is a database expert but he too is not so much familier with CUDA programming and he asked us to work on CUDA as a result.