Explain why MFLOPS are used instead of MIPS asa performance measure for supercomputers.

MFLOPS (Million Floating Point Operations Per Second) are used instead of MIPS (Million Instructions Per Second) as a performance measure for supercomputers for several reasons:

1. Nature of Operations: Supercomputers are often used for scientific calculations that require a lot of floating point operations. These operations are more complex and take more time than integer operations. MIPS measures the speed of integer operations, not floating point operations, so it's not a good measure of performance for these types of tasks.

2. Accuracy: MFLOPS is a more accurate measure of computational power for tasks that involve a lot of floating point operations. It measures the actual work done, not just the number of instructions executed.

3. Standardization: MFLOPS is a standardized measure that can be used to compare the performance of different supercomputers. MIPS can vary greatly depending on the specific instruction set used by the computer, making it less useful for comparison.

4. Complexity of Instructions: MIPS doesn't take into account the complexity of the instructions being executed. A computer executing more complex instructions may do less MIPS but more actual work. MFLOPS, on the other hand, measures the actual work done, regardless of the complexity of the instructions.

In conclusion, MFLOPS is a more accurate and useful measure of performance for supercomputers than MIPS.