A particular parallel program computation requires 100 seconds when executed on a single processor. If 40 percent of this computation is “ inherently sequential ”, then what are the theoretically best elapsed times for this program running with 2 and 4 processors, respectively ?