The tool (Oprofile or CodeAnalyst) does not provide cache event/iteration ratio directly. But the information is there.
CodeAnalyst is able to separate event samples at basic block level. The fetch_ALL event of Instruction-based sample on AMD (Family 10h) basically provdie you the information of how many times the instructions are fetched. You got the iteration count from the IBS fetch_ALL.
Then use the iteration count to calculate ratio with other Dcache events interested in.
Thank you, I will actually look into it.