See several papers, dual channel of floating-point adder algorithm, there is always such a conclusion: because in advance is the size of the tail and exchanged, tail effectively after the subtraction results are all positive, leaves out the conventional algorithms please complement step!
Is really a little don't understand why, isn't is his several operators operating together?
Really don't understand!
CodePudding user response:
This is a floating point calculations optimization algorithm, how to give directions, just watching, oneself do a play,