Recommended way to display a float* cuda array in Directx12

Accelerated Computing CUDA CUDA Programming and Performance
#1

I have a cuda array on device float *Image_dev{} result of multiple kernel calls.
It is an RGB image. In the cuda-Directx12 sample (https://github.com/NVIDIA/cuda-samples/tree/master/Samples/simpleD3D12) the result of a computation is an array of Vertex: Vertex *vertices.

Is it recommended to create a Vertex array from the float array or is there a way to create a texture that would be mapped onto a quad (rectangle)for the Directx12 part ? If yes, how should I do this ?

I’m trying to have minimum complexity in this process, to keep as much as possible the quality of the image (keep it a float*), and have it ready to be used for 10bits and HDR display in Directx.

#2

I have some results, except that the structure of the resulting image is wrong: why?

Image: https://github.com/mprevot/CudaD3D12Update/blob/master/Screenshot%20(33).png

1. there is an "all the time black" area in the top right area of the image.
2. only when I have R G B having the same value for all pixels, I get the expected result (modulo the first problem), if not I have unexpected artefacts, as if the array was not having the expected structure.

https://github.com/mprevot/CudaD3D12Update