A certain benchmark contains 195578 floating point operations, with details shown
in the table in the Figure 1 given below.
The benchmark was run on an embedded processor after compilation with optimization
turned on. The embedded processor is based on a current RISC processor
that includes floating point function units, but the embedded processor
does not include floatin point for the reaon of cost, power consumption , lack of need
for floating point by the target applications. The compiler allows the
floating point instructions to be calculated with the hardware units or using
software routines, depending on the compiler flags. The bench mark took 1.08
seconds on the RISC processor and 13.6 seconds using software on its embedded version.
Assume that the CPI(cycles per instruction) using the RISC processor was 10
and that of the embedded version of the processor was 6.
Figure 1 below : Occurences of floating point operations
Convert integer to FP (floating point) 6006
a. what is the total no. of instructions executed by both runs?
b. What is the MIPS rating for both runs?
c. On the average, how many integer instructions does it take to perform a floating point operation
in the software?