From what I have read so far, directx does a fair amount of optimization. E.g: D3D10Device->IASetInputLayout(A); D3D10Device->IASetInputLayout(B); would result in this as directx optimize by setting a dirty region. D3D10Device->IASetInputLayout(B); What about this scenario: D3D10Device->IASetInputLayout(A); DrawMeshOne(); D3D10Device->IASetInputLayout(A); DrawMeshTwo(); Will directx removes the 2nd similar IASetInputLayout() instruction? Or do I have to optimize them by myself? Reason being that each of my CObject class handles the draw code itselfs meaning CObject::Draw() does this:

CObject::Draw()
{
    IASetInputLayout(CObject::Layout);
    IASetPrimitiveTopology(CObject::Topology);
    
    for each effect pass
        DrawIndexedPrimitive();
}

Hence when many CObject::Draw() are executed, it would lead to the above scenario.

The D3D10 runtime generally (and in this case, definitely) turns redundant calls like this into a no-op before it gets to the driver. There is still some overhead in any API call, but it's probably as fast or faster than you checking yourself to see if the handles match.