1.7: Special Numbers Last updated May 31, 2022 Save as PDF 1.6: IEEE Single Precision Format 1.8: Examples of Computer Numbers Page ID96033 Jeffrey R. ChasnovHong Kong University of Science and Technology ( \newcommand{\kernel}{\mathrm{null}\,}\) (1.7.1)Smallest exponent: e=00000000, represents denormal numbers(1.f→0.f) (1.7.2)Largest exponent: e=11111111, represents ±∞, if f=0e=11111111, represents NaN, if f≠0 (1.7.3) Number Range: e=11111111=28−1=reservede=00000000=0reserved so, p=e−127 is 1−127≤p≤254−127−126≤p≤127 (1.7.4)Smallest positive normal number=1.00000000⋯⋅.0000×2−126≃1.2×10−38bin: 00000000100000000000000000000000hex: 00800000MATLAB: realmin(’single’) (1.7.5)Largest positive number=1.11111111⋯⋯⋅1111×2127=(1+(1−2−23))×2127≃2128≃3.4×1038bin: 01111111011111111111111111111111hex: 7f7fffffMATLAB: realmax(’single’) (1.7.6)Zerobin: 0000 0000 0000 0000 0000 0000 0000 0000hex: 00000000 (1.7.7)Subnormal numbersAllow 1.f → (in software)Smallest positive number =0.00000000⋯⋯0001×2−126=2−23×2−126≃1.4×10−45