directx11 – On the lookout for an recommendation on easy methods to probably enhance my compute shader

I’ve 4 rendertargets R10G10B10A2 every getting used to retailer shade and regular in a compact double RGB 555 bits format. Once I’m doing the lighting cross I am obliged to have additional rendertargets as a result of if I can learn(extract) the conventional/shade, I am unable to write again straight the lit shade to those surfaces.

So I moved to a compute shader to attain this objective (DX11 CS5). I’ve discovered that I am restricted to R32_UINT for learn/write in DX11 CS5. I’ve thus tailored my encoding/decoding to have issues work effective, though the ultimate colours are a little bit bit hugly with my encoding.

Sadly the ultimate body price is decrease than the common pixel shader even when I’ve to do 4 copyresources from the intermediate rendertargets with my earlier methodology.

I’ve applied the CS ranging from these discovered within the legacy DX11 samples. My code is like this:

CPU calls:

 //some CSbufferConstant and CSshaderresourceview settings (e.g. shadowmaps, depth textures
 gpDC11->CSSetShader(pCS, 0, 0);
 //I' mworking on 4 UAV on the similar time within the CSshader. With just one it's even slower
 gpDC11->CSSetUnorderedAccessViews(0, 4, gSRDeffered.ppUAV, (UINT*)(&gSRDeffered.ppUAV));
 gpDC11->Dispatch(960, 540, 1);//the dimensions of my display screen

GPU CS shader:

[numthreads(1,1,1)]
void CS_PostDeferred( uint3 nGid : SV_GroupID, uint3 nDTid : SV_DispatchThreadID, uint3 nGTid : SV_GroupThreadID )//solely nDTid is used actually right here
{
    uint Output;
    float2 Tex = float2(nDTid.x/960.0, nDTid.y/540.0);
    float Depth1 = txDepth1.SampleLevel(samPoint, Tex, 0).r;
    float Depth2 = txDepth2.SampleLevel(samPoint, Tex, 0).r;
    float Depth3 = txDepth3.SampleLevel(samPoint, Tex, 0).r;
    float Depth4 = txDepth4.SampleLevel(samPoint, Tex, 0).r;

    Output = UAVDiffuse0[nDTid.xy];
    if ( Depth1 < 1 ) Output = GetLColorUnPackPack_CS(Depth1, Tex, Output);
    UAVDiffuse0[nDTid.xy]=Output;
    Output = UAVDiffuse1[nDTid.xy];
    if ( Depth2 < 1 ) Output = GetLColorUnPackPack_CS(Depth2, Tex, Output);
    UAVDiffuse1[nDTid.xy]=Output;
    Output = UAVDiffuse2[nDTid.xy];
    if ( Depth3 < 1 ) Output = GetLColorUnPackPack_CS(Depth3, Tex, Output);
    UAVDiffuse2[nDTid.xy]=Output;
    Output = UAVDiffuse3[nDTid.xy];
    if ( Depth4 < 1 ) Output = GetLColorUnPackPack_CS(Depth4, Tex, Output);
    UAVDiffuse3[nDTid.xy]=Output;
}

The uint GetLColorUnPackPack_CS (float Depth, float2 UV, uint Knowledge) is the perform extracting the colour and regular as 2 float3 from the R32_UINT Knowledge param, calculating as standard lighting and shadows for the pixel after which recompacting the ultimate shade to the R32_UINT (the conventional just isn’t modified). The Depth and UV params are used to get well the place in view area wanted for the purpose lights I am utilizing.

Any recommendation welcome