My videocard doesn’t supports the global buffer g. I use the indexed_temp_array xN (scratch_buffer_IL in the samples directory) in the kernel. But I want the output array in each kernel (rough 2000 elements) and my 2 questions:
1) Can I indexing more then four elements in each output (with “dcl_output_... oN\n” I thing)?
I was found “dcl_output_RenderTargetArrayIndexed…” and “dcl_output_ViewportArrayIndexed…” in the Intermediate_Language_Specification, but I haven’t understood how to use this command yet.
2) How can I mapping the global buffer x memory after calCtxRunProgram?
When I used the calResMap standard construction I couldn’t reached the valid result.
I set the significances of x with calResMap before calCtxRunProgram and then I mapped x with calResMap standard construction and the significances of x were unchanged, though the kernel had to change it.