0 votes 0 votes Ans. D CO and Architecture co-and-architecture ieee-representation floating-point-representation + – Na462 asked Aug 6, 2018 Na462 653 views answer comment Share Follow See all 4 Comments See all 4 4 Comments reply Na462 commented Aug 6, 2018 reply Follow Share My Doubt is:- E is of 11 bits, we know that if all 11 bits are used then it represents a special number likewise if all are 0 so Emin = 00000000001 , Emax = 11111111110 so 1<=E<=2046. My doubt is why we are using all 1's, when we are given a number we define the exponent bit by adding the bias term into it for which here it will be 1023 right ? I am confused when we use bias term and when we use all 1's like here we did. Please clarify little about it 0 votes 0 votes MiNiPanda commented Aug 6, 2018 reply Follow Share We represent normalized form as (-1)sx1.Mx2E-bias. Emin = 00000000001 , Emax = 11111111110 so 1<=E<=2046. Where are we using all 1's here? Are you asking why don't we say the upper limit as 1023 as the max power of 2 is 1023? 0 votes 0 votes Na462 commented Aug 6, 2018 reply Follow Share Yes that is my doubt 0 votes 0 votes MiNiPanda commented Aug 6, 2018 reply Follow Share Okay.. it's because the Question wants to know the value of 'E' and not 'E-bias'. Max value of 2's power is 1023. So, E-bias=1023 => E=1023+bias=2046. Had it asked about the max number that can be represented( like the questions that ask for the max +ve number or the minimum -ve number that can be represented using this ieee double precision) then we had to take bias into consideration. Here it is just asking about the range of E. 1 votes 1 votes Please log in or register to add a comment.