FPU

Updated: 10/02/2017 by Computer Hope

Short for floating-point unit, the FPU is part of a computer specifically designed to perform floating-point mathematical operations. On October 30, 1994, Professor Thomas Nicely sent an e-mail indicating there was an issue with the Intel Pentium processors and how they perform floating point division.

All 120 MHz and higher Intel Pentium processors are not affected by this issue. However, if you still want to test for this problem, perform the calculation below in a calculator.

 Correct Answer
962,306,957,033 / 11,010,046 = 87,402.6282027341
 Incorrect Answer
962,306,957,033 / 11,010,046 = 87,399.5805831329
 Correct Answer
4,195,835 / 3,145,727 = 1.33382044913624100
 Incorrect Answer
4,195,835 / 3,145,727 = 1.33373906890203759

Another way of testing for this flaw is to use Microsoft Excel and enter the formula below.

=4195835-((4195835/3145727)*3145727)

When entering this formula, it returns a result of 0.

Related pages

Computer acronyms, Coprocessor, Floating-point, Floating-point coprocessor, Floating-point notations, Hardware terms