hi,

while i was anxiously waiting for the NodeKO results i was taking a look
at CUDA http://www.nvidia.com/object/cuda_home.html .

I have a MacBook Pro with an Nvidia 330m, which has CUDA support,
downloaded the SDK and docs and was preparing to take a look at it, but
i'm a C newb.


I found a page about GPU computing on Wikipedia which explains the
architecture on http://en.wikipedia.org/wiki/CUDA .

All i know of the node.js internals is what i seen on the node video
conf talk, and if my memory doesn't fail it talks about node having on
process and a single code execution stack? The wikipedia page states:

"Threads should be running in groups of at least 32 for best
performance, with total number of threads numbering in the thousands.
Branches in the program code do not impact performance significantly,
provided that each of 32 threads takes the same execution path; the SIMD
execution model becomes a significant limitation for any inherently
divergent task (e.g. traversing a space partitioning data structure
during raytracing)."

is it worth exploring, what do you think? If so, i call dibs on a
project name, nodeGT ;P



best

-- 
http://zemanel.eu
http://github.com/zemanel
http://pt.linkedin.com/in/josemoreira
http://djangopeople.net/josemoreira
irc://[email protected]

-- 
v8-users mailing list
[email protected]
http://groups.google.com/group/v8-users

Reply via email to