| 摘要 |
PURPOSE:To detect and correct arithmetic errors by computing data having two different radixes as a n-notation number and an m-notation number and respectively comparing remainders obtained by subtracting (n) and (m) from arithmetic results. CONSTITUTION:Input data are mutually added by an adder 1a, one input data are applied to respective remainder computing elements 2a, 2c to find out respective remainders obtained by respectively subtracting (n) and (m) from the input data and the other data are applied to respective residual computing elements 2b, 2d to find out respective remainders obtained by respectively subtracting (n) and (m) from the input data. Then the outputs of the computing elements 2a, 2b are added by an adder 1b to find out a remainder by subtracting (n) from the added result and the outputs of the computing elements 2c, 2d are added by an adder 1c to find out a remainder by subtracting (m) from the added result. Then respective remainder computing elements 2e, 2f find out remainders by respectively subtracting (n) and (m) from the output of the adder 1a and comparators 3a, 3b respectively compare the outputs of the computing elements 2e, 2f and the adders 1b, 1c. Then a corrector 5 corrects the output of the adder 1a based upon an error bit specified by an error deciding device 4. |
