This API may do it, meanwhile, a single call to executeCommandsInBuffer may not outperform the loop of drawIndexedPrimitives. Take a look at this article, link: https://tellusim.com/metal-mdi/ Anyhow, thanks all the same.

My understanding(not sure right or not): In the indirect command buffer, one command for sphere, one command for cube, so actually executeCommandsInBuffer executes two commands, two draw calls. For ExecuteIndirect , it may go like that: CPU calls it once, sends data to GPU, then GPU draw the two objects in two separate commands. One draw call on CPU, two draw steps on GPU

Thank you. May be I did not specify clearly, what I want is: for example, there are one sphere and one cube, I could just call DX 12 API ExecuteIndirect once(one draw call) to render the two objects on Windows platform. Now is it possible to call some API to do the same thing on iOS platform?