#include "Include/PlatformDefines.hlsl"

#pragma kernel CSInstancedBufferToTextureKernel
#pragma kernel CSBufferToTextureCrossFadeKernel

RWStructuredBuffer<float4x4> gpuiInstanceData;
ConsumeStructuredBuffer<uint> gpuiTransformationMatrix;
RWTexture2D<float4> gpuiTransformationMatrixTexture;
RWStructuredBuffer<uint> argsBuffer;
uniform uint argsBufferIndex;
uniform uint maxTextureSize;

RWStructuredBuffer<uint4> gpuiInstanceLODData; // lodNo - shadowLodNo - cfLodNo - cfLevel 
RWTexture2D<float4> gpuiInstanceLODDataTexture;
uniform uint bufferSize;

[numthreads(GPUI_THREADS, 1, 1)]
void CSInstancedBufferToTextureKernel(uint3 id : SV_DispatchThreadID)
{
    if (id.x >= argsBuffer[argsBufferIndex])
        return;
    uint indexX = id.x % maxTextureSize;
    uint indexY = floor(id.x / float(maxTextureSize));

    uint instanceId = gpuiTransformationMatrix.Consume();
    
    float4x4 transformData = gpuiInstanceData[instanceId];
    gpuiTransformationMatrixTexture[uint2(indexX, 0 + indexY * 4)] = transformData._11_12_13_14; // row0
    gpuiTransformationMatrixTexture[uint2(indexX, 1 + indexY * 4)] = transformData._21_22_23_24; // row1
    gpuiTransformationMatrixTexture[uint2(indexX, 2 + indexY * 4)] = transformData._31_32_33_34; // row2
    gpuiTransformationMatrixTexture[uint2(indexX, 3 + indexY * 4)] = float4(transformData._41_42_43, instanceId); // row3
}

[numthreads(GPUI_THREADS, 1, 1)]
void CSBufferToTextureCrossFadeKernel(uint3 id : SV_DispatchThreadID)
{
    if (id.x >= bufferSize)
        return;
    
    gpuiInstanceLODDataTexture[uint2(id.x % maxTextureSize, floor(id.x / float(maxTextureSize)))] = gpuiInstanceLODData[id.x];
}