2 votes 2 votes Mantissa is a pure fraction in signed magnitude form .What is the reprentation of decimal number 12.6255*102 12.6255*23 with Normalization and Without Normalization CO and Architecture floating-point-representation + – Dulqar asked Jan 23, 2017 Dulqar 1.5k views answer comment Share Follow See all 16 Comments See all 16 16 Comments reply Pavan Kumar Munnam commented Jan 23, 2017 reply Follow Share normalization is nothing but representing the floating point in 1.xx this way it is normalized and any other way it is not normalized 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share But how to store other way numbers ? 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share Just tell me one thing If the number is Binary then 1111.1111 *2^2 How to store this without normalization technique ? 0 votes 0 votes Pavan Kumar Munnam commented Jan 23, 2017 reply Follow Share we cannot store unnormalized in ieee754 , we normalize just to prevent a number for having two representations, we restrict that by having the format as 1.xx otherwise you will have two representation for same number so either it should go as 1.xx or 0.1xx like this i think 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share @dulkar why we use normalisation? 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share Acctually I dont know . I have seen questions asking Normalized and Without normalized represeantation of a number . And Im geting confused .!! :( sudsho can u pls help me understand this 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share see suppose u want to represent 4.5.. in binary it is 100.1 right so to represent it in floating point u can write it like 0.1001 * 23 also u can write it like 0.01001*24..this also represent 4.5..means u are having multiple representation for a single no... to avoid this we use normalisation....we represent 4.5 only using 0.1001 * 23 ...no other representation so now ur question is to represent without normalisation...just cconvert the no into binary and put everything in mantessia and exponent field as it is..no biuasing no normalisation anything.. but this is not the only way to represent a no without normalisation...we can have multiple ways..so asking to represent the no without normalisation is nothing but normal binary representation where u also say where the floating point is..like i can represent 4.5 also like(1001,3)..so there can be multiple representations and hence this question was baseless...i hope u got the concept :) 1 votes 1 votes sudsho commented Jan 23, 2017 reply Follow Share in normalisation we have exact representation for a no...u can use either explicit where we put radix point after just before most significant 1 or implicit normalisation where we put radix point just after most significant 1..i have solved one such question few days back...it was yours only...where pc also answered considering decimal..remember? 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share @sudsho Thank you. so now ur question is to represent without normalisation...just cconvert the no into binary and put everything in mantessia and exponent field as it is..no biuasing no normalisation anything.. Can u just show one exaple for this one also ? For Example 1111.1111 *2^2 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share 1111.1111 *2^2 =0.11111111 *26 put 11111111 in mantesia..and 6 in exponent ..done! 1 votes 1 votes Dulqar commented Jan 23, 2017 reply Follow Share If it were Normalized then exponent value would be 70? 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share no... 1111.1111 *2^2 is represented as 0.11111111 *26 also i can represent it 0.011111111 *26 or (11111111,6) and various other...but in normalisation only one representation is there a\with biasing..(note that m representing 6 as it is in exponent without biasing) 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share yes...exponent will be bias+6 means 70 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share One last question . In question if it is given Without Normalization and Rounding of . Then also we consider normalised approach while solving ? 0 votes 0 votes sudsho commented Jan 23, 2017 reply Follow Share they will not ask without normalisation...we can have various representations in such case.. if asked then why u would consider normalised approach?...just convert into binary thats it..and state where is the radix point... in one line without normalisation is nothing but how u represent any fractional decimal in binary thats it! 0 votes 0 votes Dulqar commented Jan 23, 2017 reply Follow Share https://gateoverflow.in/109784/made-easy-cbt-2017-q13 See this one. It is confuses me :( 0 votes 0 votes Please log in or register to add a comment.