Date
21 - 23 of 23
State of OCIO development re: CPU/GPU mismatches
Dithermaster <dither...@...>
> Dennis, is your fork containing the OpenCL code path available to browse currently? Nathan, no, but I could put something together for you if you'd like to review it. ///d@ On Mon, Sep 12, 2016 at 8:15 PM, Nathan Rusch <natha...@...> wrote:
|
|
Dithermaster <dither...@...>
Nathan and others interested, I've attached a PDF describing our OpenCL extensions to OpenColorIO. An accurate OpenGL path would do something very similar with multiple shader parameters, and would return GLSL instead of an OpenCL kernel. I am happy to work closely with anyone who wants to create the accurate OpenGL path starting from these changes, but I'm not an OpenGL expert so you'd need to handle the OpenGL side of things. Sincerely, ///d@ Dennis Adams On Tue, Sep 13, 2016 at 10:45 AM, Dithermaster <dither...@...> wrote:
|
|
Sean Cooper <se...@...>
I think it would be very interesting to take a look at an OpenCL OCIO implementation. It seems like: #304, #396, + Kevin's CDL parse, would be easy first steps to get these creaky joints moving again. I'll have more time next week to investigate the other pull requests, and catch up on the test suite. @Malcolm: do know the status of the DNeg OCIO pull that has been referred to? And if there are still people there interested in pushing to the public repo? On Tue, Sep 13, 2016 at 9:37 AM, Dithermaster <dither...@...> wrote:
|
|