Computing using Graphics Cards
14 years ago
| 41 interested
Graphics cards(GPUs), once made purely for games, are now being increasingly used for general purpose computation. (see http://gpgpu.org/about for some more details)
GPUs can outperform CPUs for a large number of computations, while costing much less. Programming GPUs is not as simple as programming CPUs. However, the results can many times outweigh the effort.
GPUs are an attractive platform for compute acceleration, and are becoming ubiquitous.
This session aims to discuss
– advantages of GPUs
– usage areas
– comparison of GPU & CPU architectures
– programming for GPUs
– available tools…
Share this session:
Shree, Session sounds quiet interesting.
Are you planing to do any demo ?
I wish to include one more point here, when not use GPU and use CPU :).
I can make a demo. The problem I face is : I don’t have a laptop capable of running CUDA apps. If somebody has a decent laptop with a GPU, we can make that happen.
I will surely talk about when to use the CPU & when to use the GPU. The short answer, for now, is : use the CPU for serial jobs & use the GPU for massive parallel processing.
Sounds good. I think we should able to get at least on laptop with CUDA based GPU at BCB9.
@Viral : That would be fantastic! I can show guys several characteristics if a GPU is available. More practical than theory works every time …
I was looking something like this from lot of days.
I’m having a laptop with nVedia Card and will try to get CUDA installed before the session.
Looking forward to meet you in session.
@1989gaurav: You’re my man !
[…] up of sessions in various technical domain. I was most interested in Shree kumer‘s session on using GPU for computing, which turned out to be quite interesting. Finally, I seen some working demo of CUDA application […]