You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
struct SpdConstants
{
int mips;
int numWorkGroupsPerSlice;
int workGroupOffset[2];
};
struct SpdLinearSamplerConstants
{
int mips;
int numWorkGroupsPerSlice;
int workGroupOffset[2];
float invInputSize[2];
float padding[2];
};
To ffx_spd.h under a A_CPU define. That would make it more convinient of using just
FWIW, I can see why this remains in the SPD samples. In fact, I would go one step further and suggest that all the CPU code be separated altogether. The SPD constants are bound to be sourced very differently in different engines since you may not need the workgroup offset support, multiple-slice operation, and whatnot. Not to mention that you might want to pack this data along with other data that's persistently mapped in the frame or something (it's not uncommon for framebuffer metadata to be stored somewhere and reused by other postprocessing work). As a final consideration, different renderers/engines have different ways to "glue" data on the host and device.
I think it would be better to move
To ffx_spd.h under a A_CPU define. That would make it more convinient of using just
In the implementation file. Currently they are hidden in example project file.
The text was updated successfully, but these errors were encountered: