This is all very much a work in progress now.

Thanks, and yeah, the whole GPU effort could really use more hands on it.

On Oct 22, 2020, at 12:13 PM, Sam Assadian <sam@...> wrote:

Fantastic!

While it's still very early for us right now, our exploring is moving pretty fast (thanks for the great support btw). If you guys need help by the time our devs are moving to the next level, just know that we would be happy to help.

Thanks!

On Thu, Oct 22, 2020 at 9:01 PM Larry Gritz <lg@...> wrote:

Streaming? If you mean the on-demand loading, then yes.

On Oct 22, 2020, at 11:57 AM, Sam Assadian <sam@...> wrote:

Hey guys! just chiming in here and sorry if that has been mentioned already: are there any plans to extend the texture system to support streaming in OptiX OSL?

Cheers,

Sam

Sam Assadian CEO / Co-Founder +33 972 452 846

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed.

On Thu, Oct 22, 2020 at 8:46 PM Anders Langlands <anderslanglands@...> wrote:

I get the sense that the intention is there to make it a complete implementation, it just isn’t finished yet. Which is what I’m planning on doing next: starting with the matrix functions. 

On Fri, 23 Oct 2020 at 04:56, <bruiz@...> wrote:

I would also add that the provided rend_lib.cu doesn't define some functions required when evaluating textures (for example osl_texture_set_swrap_code(...) and osl_texture_set_twrap_code() were missing). So, as far as I can tell, one has to modify this lib to support such features, and it thus becomes impractical to use the one provided in the examples. To me, providing the rend_lib as an llvm bitcode to OSL seems completely unnecessary.
It would be interesting to know why this choice was made in the first place.

Cheers,
Benoit R.

--

Larry Gritz

lg@...

--

Larry Gritz

lg@...