[D3D11] DrawInstancedIndirect slower than DrawInstanced with the same parameters on R9 290
I'm drawing a lot of geometries in one draw call using the DrawInstanced function, and I wanted to add GPU culling.
My first step was to use the DrawInstancedIndirect function, with an indirect buffer containing the exact same parameters use in the DrawInstanced function (since for now I have a fix number of these objects).
I noticed that when using the indirect function, the draw call is slower (~1.0ms) than when I use the regular function (~0.6ms).
I couldn't find the issue in my code, so I tried it on a computer with a NVIDIA GT 630m, and then when switching from one function to the other the timing is always the same (~7.5 ms).
Is there a reason for DrawInstancedIndirect being slow on a R9 290 ? I'm using the 14.9 drivers, should I try the latest beta drivers or maybe some older ones ?