看板 Grad-ProbAsk 關於我們 聯絡資訊
※ 引述《mickeyha (M*schief)》之銘言: : 張凡上冊p.168練習(下面那一題) : The IEEE 754 floating-point standard specifies 64-bit double precision : with a 53-bit significand ( include the implied 1 ) and an 11-bit exponent. : IA-32 offer an extended precision option with a 64-but significand and a 16-bit : exponent. : 想請問的是第二小題: : (2) What is the range of numbers that can be represented by the extended : presion option? : 以最小來說,我想的是 1.0000....000 * 2^(-32767-64) = 1.0 * 2^(-32831) : 二 進 位 ^^^^^^^^^^弄錯點1 : 也就是十進位的 2.0 * 10 ^ (-32831*0.3) = 2 * 10^(-9849) ^^^^^^^^^^^^^^弄錯點2 1.0 * 2^(-32831) = 2.0 * 2^(-32832) = 2.0 * 10 ^ (-32832*0.3) : 但答案是 2.0 * 10(-9864) : 十進位 : 想請問我是哪裡弄錯了呢...謝謝":) 我認為應該是 1.000....000 * 2^(1-32767) = 1.0 * 2^(-32766) (P.S 其實這也可看成是10進位的"值"喔) 1.0 * 2^(-32766) = 10^(-32766*0.301) ≒ 10^(-9863) = 2.0 * 10^(-9864) -- ↑↑↑ (((──╮ o O ○ ███◣ ◢██◣ 濱口組長眉頭一皺, ██████ \ ㄟˇㄏ / ██████ ⊙..⊙ 覺得插到的魚並不單純。 ████ /◎||||\╰┼╯ cynic5566 -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 140.113.138.24
mickeyha:謝謝您:) 12/10 01:19