Consider the case where a program consists of 30% integer instructions, 50% floating-point instructions and 20% memory access operations and executes 90 billion instructions. Let a machine take 260 ps to execute an integer instruction, 638 ps to execute a floating-point instruction and 8 ns to complete a memory access operation. Assume that we can overclock the processor arbitrarily, thus speeding up the integer and floating-point performance by an arbitrary factor. What is the fastest we could finish computing this program just by overclocking the processor?