//TEST(compute):COMPARE_COMPUTE_EX:-slang -compute -shaderobj -output-using-type
//TEST(compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute -shaderobj -output-using-type
//TEST(compute):COMPARE_COMPUTE_EX:-cuda -compute -shaderobj -output-using-type

//TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0 0 0 0], stride=4):out,name=outputBuffer
RWStructuredBuffer<float> outputBuffer;

typedef DifferentialPair<float3> dpfloat3;
typedef float3.Differential dfloat3;

typedef DifferentialPair<float2> dpfloat2;
typedef float2.Differential dfloat2;

[BackwardDifferentiable]
float3 test_simple(float3 x, float3 y)
{
    return x + y;
}

[BackwardDifferentiable]
float test_swizzles(float3 x, float3 y)
{
    return x.y + y.x;
}

[BackwardDifferentiable]
float3 test_constructor(float3 x, float3 y)
{
    return float3(x.y + y.x, y.z, x.z);
}

[BackwardDifferentiable]
float3 test_complex_arith(float3 x, float2 y)
{
    float2 t = float2(x.x, y.x) + 2.0 * float2(x.z, y.y);
    return float3(
        (2.0f * x.z + y.y * 4.0f - t.y), 
        (3.0f * t + y).x,
        t.x);
}

[numthreads(1, 1, 1)]
void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
{
    {
        dpfloat3 dpx = dpfloat3(float3(2.0, 3.0, 4.0), float3(0.0, 0.0, 0.0));
        dpfloat3 dpy = dpfloat3(float3(1.5, 2.5, 3.5), float3(0.0, 0.0, 0.0));

        __bwd_diff(test_simple)(dpx, dpy, dfloat3(1.0, 2.0, 3.0));
        outputBuffer[0] = dpx.d.y; // Expect: 2
        outputBuffer[1] = dpy.d.y; // Expect: 2
    }

    {
        dpfloat3 dpx = dpfloat3(float3(2.0, 3.0, 4.0), float3(0.0, 0.0, 0.0));
        dpfloat3 dpy = dpfloat3(float3(1.5, 2.5, 3.5), float3(0.0, 0.0, 0.0));

        __bwd_diff(test_swizzles)(dpx, dpy, 2.3);
        outputBuffer[2] = dpx.d.y; // Expect: 2.3
        outputBuffer[3] = dpy.d.x; // Expect: 2.3
        outputBuffer[4] = dpy.d.y; // Expect: 0.0
    }

    {
        dpfloat3 dpx = dpfloat3(float3(2.0, 3.0, 4.0), float3(0.0, 0.0, 0.0));
        dpfloat3 dpy = dpfloat3(float3(1.5, 2.5, 3.5), float3(0.0, 0.0, 0.0));

        __bwd_diff(test_constructor)(dpx, dpy, float3(1.0, 1.5, 2.0));
        outputBuffer[5] = dpx.d.y; // Expect: 1.0
        outputBuffer[6] = dpy.d.x; // Expect: 1.0
        outputBuffer[7] = dpy.d.z; // Expect: 1.5
    }

    {
        dpfloat3 dpx = dpfloat3(float3(2.0, 3.0, 4.0), float3(0.0, 0.0, 0.0));
        dpfloat2 dpy = dpfloat2(float2(1.5, 2.5), float2(0.0, 0.0));

        __bwd_diff(test_complex_arith)(dpx, dpy, float3(1.0, 1.5, 2.0));
        outputBuffer[8] = dpx.d.y; // Expect: 0.0
        outputBuffer[9] = dpy.d.x; // Expect: -1.0 + 1.5 = 0.5
        outputBuffer[10] = dpx.d.z; // Expect: 2.0 + 9.0 + 4.0 = 15.0
    }

}