Re: Compiling OpenShadingLanguage under Windows


Wormszer <worm...@...>
 

After looking at ATI's open cl I think its a little bit overkill. From initial look i think we would have to create kernels for each math function we wanted as well as create an open cl context to make use the the language features defined in the Open CL spec.

For doing a gpu version of a shader this might work but as a stand-in for the C99 math i don't think it would be a good solution.

Jeremy


On Mon, Jan 18, 2010 at 8:58 PM, Wormszer <worm...@...> wrote:
I tried the first link you mention but i end up with a different set of errors.

What did you do to get it working? I will play with it some more and see.

I got the defines enabled by defining the _USE_MATH_DEFINES like Chris mentioned and mentioned in the second link.

I agree on trying to use something like open CL. Do they provide CPU implementations for all the math functions?
But to take advantage of the real gpu speed up we would need to do it somewhere else where things can be batched.

I hope once i get this building to dig in and figure out what is going on in the shader execution and language parsing. With my goal trying to help with getting some kind of GPU based shader preview implementation going as mentioned in Larry's introduction.
Open CL I would think should be a good language to shoot towards as I am under the impression its supposed to replace nvidias cuda and ati's brook stream stuff. Or at least both vendors will support it.

Jeremy

On Mon, Jan 18, 2010 at 6:19 PM, Oleg <ode...@...> wrote:
Hi Jeremy,

I have installed the latest cygwin version and was able to process all
flex/bison files. The cygwin flex version can handle "/" as path
separator, too :-) Apart from that I had the same problems compiling
"oslinfo", "oslquery", "oslcomp" and "exec" projects. I have found c99
macro definitions

here    http://lists.boost.org/Archives/boost/2005/12/97593.php

and

here    http://www.gamedev.net/community/forums/topic.asp?topic_id=127944

Maybe we can use CUDA or OpenCL (which is better, since available on
NVIDIA/ATI) for all missing math functions. Please see

- http://developer.download.nvidia.com/compute/cuda/2_3/toolkit/docs/NVIDIA_CUDA_Programming_Guide_2.3.pdf,
page 120
- http://www.khronos.org/registry/cl/specs/opencl-1.0.43.pdf?q=specification,
page 168

Regards,
Oleg


On 18 Jan., 05:51, Wormszer <wo...@...> wrote:
> My generated code issues were from VS having not rebuilt the files with the
> new flex.
>
> So now I am just left with a bunch of missing math functions.
> It looks likes visual studios implementation of math. or cmath is missing a
> lot of the functions.
>
> Things like
> log2f, logbf, truncf, etc
> and defines like M_E, M_1_PI, M_PI_2 etc.
>
> See if you run into the same thing and then im not sure where to go from
> here.
>
> I guess we should get some feedback from others, maybe they are available
> somewhere else, or I could be missing something.
>
> Jeremy
>

--
You received this message because you are subscribed to the Google Groups "OSL Developers" group.
To post to this group, send email to osl...@....
To unsubscribe from this group, send email to osl...@....
For more options, visit this group at http://groups.google.com/group/osl-dev?hl=en.





Join osl-dev@lists.aswf.io to automatically receive all group messages.