How to Compute MFLOPS in Early Computers
The term MFLOP, or megaFLOP, refers to how many millions of floating point operations per second a computer can perform. Early super computer performance was measured in FLOPS and MFLOPS until 1988, when the Cray Y-MP, the first supercomputer to sustain over one gigaFLOP performance, was introduced. As of 2011, supercomputer performance is measured in petaFLOPS.
Other People Are Reading
Instructions
-
-
1
Load the early computer with a program that adds one to itself repeatedly.
-
2
Start the program and let it run for 60 seconds.
-
-
3
Stop the program. Read the number that it outputs, being the total number of floating point instructions that it executed.
-
4
Divide that number by 60 to get the result in FLOPS.
-
5
Divide the result in FLOPS by 1,000 to get the final result in MFLOPS.
-
1
Tips & Warnings
If you have the FLOPS measurement of an early computer, divide it by 1,000 to convert it to MFLOPS. If you have a measurement in GFLOPS, multiply it by 1,000 the get a measurement in MFLOPS.
You can determine the FLOPS rating of your own modern computer by using a tool such as the online Linpack Benchmark test, which uses Java.
