By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
455,462 Members | 1,516 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 455,462 IT Pros & Developers. It's quick & easy.

How is parallel stuff going

P: n/a
Hi,
there are some concurrency frameworks available for .net from microsoft,
which takes advantage of the cpu. But recently GPUs have broken the 1 tera
flop barrier and are really ideal for parallel computing. For example
Nvidia's CUDA allows us to program the GPU. My question: is there any effort
going inside microsoft that would give us (the devs) some framework, which
wraps up the functionality of CUDA (or similar technology from amd/ati) and
we program through those special APIs that gets our parallel code executed
inside the gpu rather then tha cpu.

regards,

...ab
Jul 9 '08 #1
Share this Question
Share on Google+
3 Replies


P: n/a
Abubakar,

I imagine the answer is not yet. My understanding is that there is not
a standardized way to access the GPUs yet, so having any solution would be
provider-specific, and that's not really going to help in creating a
framework which will address a number of different GPUs.

I imagine that OpenCL is a step in this direction, and it's something
that you might want to consider using in .NET, if you have hardware that
supports it.

--
- Nicholas Paldino [.NET/C# MVP]
- mv*@spam.guard.caspershouse.com

"Abubakar" <q@y.comwrote in message
news:OM**************@TK2MSFTNGP02.phx.gbl...
Hi,
there are some concurrency frameworks available for .net from microsoft,
which takes advantage of the cpu. But recently GPUs have broken the 1 tera
flop barrier and are really ideal for parallel computing. For example
Nvidia's CUDA allows us to program the GPU. My question: is there any
effort going inside microsoft that would give us (the devs) some
framework, which wraps up the functionality of CUDA (or similar technology
from amd/ati) and we program through those special APIs that gets our
parallel code executed inside the gpu rather then tha cpu.

regards,

..ab


Jul 9 '08 #2

P: n/a
Abubakar wrote:
there are some concurrency frameworks available for .net from microsoft,
which takes advantage of the cpu. But recently GPUs have broken the 1 tera
flop barrier and are really ideal for parallel computing. For example
Nvidia's CUDA allows us to program the GPU. My question: is there any effort
going inside microsoft that would give us (the devs) some framework, which
wraps up the functionality of CUDA (or similar technology from amd/ati) and
we program through those special APIs that gets our parallel code executed
inside the gpu rather then tha cpu.
You can always build a Win32 DLL from some C code and access
that from C#.

Google finds stuff like:

http://research.microsoft.com/resear...report&id=1040
http://research.microsoft.com/Resear...d/Details.aspx

and

http://brahma.ananthonline.net/

which you could try out.

Arne
Jul 10 '08 #3

P: n/a
Hello Abubakar,

I believe F# in a step in this direction. I think functional approach is
much clear than imperative for these parallel stuff; and as soon as we don't
have any mainstream functional language from Microsoft I wouldn't expect
anything really usefull.

Kind regards,
Andrei Varanovich [C# MVP]
Hi,
there are some concurrency frameworks available for .net from
microsoft,
which takes advantage of the cpu. But recently GPUs have broken the 1
tera
flop barrier and are really ideal for parallel computing. For example
Nvidia's CUDA allows us to program the GPU. My question: is there any
effort
going inside microsoft that would give us (the devs) some framework,
which
wraps up the functionality of CUDA (or similar technology from
amd/ati) and
we program through those special APIs that gets our parallel code
executed
inside the gpu rather then tha cpu.
regards,

..ab

Jul 13 '08 #4

This discussion thread is closed

Replies have been disabled for this discussion.