Re: OpenCL

[Benjamin Tolputt <btolputt@...>]

Rob Kendrick wrote:
> I'm having trouble imagining a situation where there'd be a significant
> performance boost from running an interpreted general-purpose language
> on your GPU, though.  They're really only any good at highly parallel
> workloads.

I agree - general purpose Lua scripts are not likely to run well on a
SIMD architecture like a GPU. What might be nice though is a Lua
interface to the functionality of OpenCL... once they have a release
ready for the masses.

As a background for my desire, I'm looking at evolutionary processes for
agents in a virtual world. I am trying to combine evolutionary/genetic
programming of Lua "control" code using inputs from layered Restricted
Boltzmann Machines (a specialised and useful artificial neural network
architecture). The GPU is perfect for the training/evaluation of the
Boltzmann Machine layers weights/values (as they can be loaded into
textures and processed with a fragment shader) with the Lua control
agents being evolved to use these inputs in the standard fashion. If
only I could find someone willing to pay me for the time spent doing that!

-- 
Regards,

Benjamin Tolputt
Analyst Programmer
Email: btolputt@bigpond.net.au

This email and any files transmitted with it are confidential to the
intended recipient and may be privileged. If you have received this
email inadvertently or you are not the intended recipient, you may not
disseminate, distribute, copy or in any way rely on it. Further, you
should notify the sender immediately and delete the email from your
computer. Whilst we have taken precautions to alert us to the presence
of computer viruses, we cannot guarantee that this email and any files
transmitted with it are free from such viruses.