tags:

views:

132

answers:

4

Hi,

I've a university project in which I should find and implement a proper development topic for Azure. In Azure there are Web Roles and Worker Roles and I would like to find something, that burdens relatively big load on both roles.

(web role is for serving the users by webpages, worker roles are for background processing data that will be displayed via the web role)

One of the goals is to show up the huge computing power of the cloud by measuring. My problem is, I don't know any sophisticated algorithms that would make the CPUs sweat. So what would be the thing that is not too complex for implement but complex enough to be computed?

Any ideas appreciated!

+3  A: 

Cracking password is always a fun exercice with a lot of computing power.

That's how I "learned" CUDA!

Also, the CUDA page is full of ideas

http://www.nvidia.com/object/io_1260777476567.html

there are also examples in the CUDA SDK

http://developer.download.nvidia.com/compute/cuda/sdk/website/samples.html

Eric
+2  A: 

Global illumination such as photon mapping or radiosity is a good CPU and memory intensive task. At work we have it running distributed otherwise it takes over 24 hours on a single PC to calculate for a large scene.

zebrabox
+1  A: 

You could write some machine learning code. For example Naive Bays Text Classification works well and there are lots of datasets out there, including some huge ones

Chad Okere
A: 

You can try a video coding algorithm, I think it is one of the most consuming task in computer processing. You don't have to implement by yourself a video codec. You can find some implementation in the web, called reference codec, for H.263 and H.264 for example.

Maurizio Reginelli