Intel-compatible processors also support an “extended precision” floating-point format w...

Intel-compatible processors also support an “extended precision” floating-point format with an 80-bit word divided into a sign bit, k = 15 exponent bits, a single integer bit, and n = 63 fraction bits. The integer bit is an explicit copy of the implied bit in the IEEE floating-point representation. That is, it equals 1 for normalized values and 0 for denormalized values. Fill in the following table giving the approximate values of some “interesting” numbers in this format:

This format can be used in C programs compiled for Intel-compatible machines by declaring the data to be of type long double.However, it forces the compiler to generate code based on the legacy 8087 floating-point instructions.The resulting program will most likely run much slower than would be the case for data type float or double.