To perform millions of Floating point Operations per Second
The CPU’s ability to process floating-point arithmetic
Suitable for measuring vector machines
To correctly reflect the performance of the computer, each floating-point operation is multiplied by a regularized value and then evaluated for the MFLOPS