calculate amat if it is known that on average, 80% of l1 cache access latency is overlapped with computation, 40% of l2 cache access latency is overlapped with computation and 10% of l2 miss latency is overlapped with computation. assume that on average, 3 memory references are serviced simultaneously for l1 cache accesses, 2.5 memory references are serviced simultaneously for l2 cache accesses, and only 1.5 for l2 cache misses ?