A certain program uses ALU, Load, Store and Branch instructions. We propose improvements that will reduce the CPI of the Load (from 5 to 2), Store (from 3 to 2) and Branch (from 2 to 1) instructions. After improving the machine that will be running this program, the Load instructions use 30% of the CPU execution time, and the Store and Branch instructions use 16% of this time each. What is the total speedup that we achieve with these improvements?