Binary Multiplication
Note: With two’s complement multiplication, the leading “1” in the second multiplicand is a
sign bit. If the sign bit is “1”, then take the 2’s complement of the first multiplicand.
Additionally, each partial product must be sign-extended for correct computation.
Note: All of the above questions except the final one are addressed in this module. The last
question may have several answers:
• Store the lower accumulator to memory. What problem is apparent using this
method in this example?
• Store the upper accumulator back to memory. Wouldn't this create a loss of
precision, and a problem in how to interpret the results later?
• Store both the upper and lower accumulator to memory. This solves the above
problems, but creates some new ones:
− Extra code space, memory space, and cycle time are used
− How can the result be used as the input to a subsequent calculation? Is such a
condition likely (consider any “feedback” system)?
From this analysis, it is clear that integers do not behave well when multiplied. Might some other
type of number system behave better? Is there a number system where the results of a
multiplication are bounded?
C28x - Numerical Concepts & IQmath 8 - 7
To Next Page
Loading...
Need help?
General
