The real logarithmic number system, which represents a value with a sign bit and a quantized logarithm, can be generalized to create the complex logarithmic number system, which replaces the sign bit with a quantized angle in a log/polar coordinate system. Although multiplication and related operations are easy in both real and complex systems, addition and subtraction are hard, especially when interpolation is used to implement the system. Both real and complex logarithmic arithmetic benefit from the use of co-transformation, which converts an addition or subtraction from a region where interpolation is expensive to a region where it is easier. Two co-transformations that accomplish this goal are introduced. The first is an approximation based on real analysis of the subtraction logarithm. The second is based on simple algebra that applies for both real and complex values and that works for both addition and subtraction.

The real logarithmic number system, which represents a value with a sign bit and a quantized logarithm, can be generalized to create the complex logarithmic number system, which replaces the sign bit with a quantized angle in a log/polar coordinate system. Although multiplication and related operations are easy in both real and complex systems, addition and subtraction are hard, especially when interpolation is used to implement the system. Both real and complex logarithmic arithmetic benefit from the use of co-transformation, which converts an addition or subtraction from a region where interpolation is expensive to a region where it is easier. Two co-transformations that accomplish this goal are introduced. The first is an approximation based on real analysis of the subtraction logarithm. The second is based on simple algebra that applies for both real and complex values and that works for both addition and subtraction.

D.M. Lewis (ibid., vol.39, no.11, p.1325-36, Nov. 1990) recently described an architecture for performing addition and subtraction in the logarithmic number system using a Taylor series approximation and ROM tables. Here it is shown that a 25% improvement can be achieved in the table used to implement addition. This improvement is significant when Lewis' technique is used to implement variations of the logarithmic number system which do not require the subtraction logarithm.>

Various features found in standard floating point arithmetic (IEEE 754) are examined in light of their appropriateness for sign/logarithm arithmetic. The emphasis is on a 32-b word size comparable to IEEE 754 single precision, although other word sizes are possible. A multilayer sign/logarithm format is considered. The lowest layer, similar to previous implementations, would provide only normalized representations but would not provide representations for zero, denormalized values, infinities, and NaNs. The highest layer would provide most of the features found in IEEE 754, including zeros, denormalized values, infinities, and NaNs. Novel algorithms for implementing logarithmic denormalized arithmetic are presented. Simulation results show that the error characteristics of the proposed logarithmic denormalized arithmetic algorithms are similar to those of the denormalized floating point arithmetic in IEEE 754.>

A number system that offers advantages in some situations over conventional floating point and sign/logarithmic number systems is described. Redundant logarithmic arithmetic, like conventional logarithmic arithmetic, relies on table lookups to make the arithmetic unit simpler than an equivalent floating point unit. The cost of 32 bit subtraction in a redundant logarithmic number system is lower than previously published logarithmic subtraction methods. The total memory requirement for a 29-bit redundant logarithmic unit is 16 K words compared to 22 K words by the best previously published conventional sign logarithm unit, assuming similar addition techniques are employed. A redundant logarithmic number system can be implemented with online arithmetic, which would be impractical for a conventional sign logarithm number system. The disadvantages of redundant arithmetic are typical of redundant number systems. First, the redundancy doubles the storage requirements for data values. Second, the representation can become ill-conditioned, especially as a result of iterated multiplications. Third, division and square root operations are more difficult to implement in redundant logarithmic arithmetic.>

A new number system that offers advantages over conventional floating-point and sign/logarithm number systems is described. Called redundant logarithmic arithmetic, it relies, like conventional logarithmic arithmetic, on table lookups to make the arithmetic unit simpler than an equivalent floating-point unit. The cost of 32-b subtraction in a redundant logarithmic number system is lower than that of previously published logarithmic subtraction methods. Another advantage of a redundant logarithmic number system is that a single arithmetic unit can use the same hardware to add, subtract, or multiply in similar times.>

Logarithmic number systems are an attractive method of implementing high-speed digital signal-processing systems with a word size of about 14 bits. Larger word sizes pose problems because of the address limitations of the ROMs needed for logarithmic addition. M.G. Arnold (1982) gave several interpolation techniques that can double the precision, but these techniques require the use of a fixed-point multiplier. F.J. Taylor (1983) proposed a modified linear interpolation that could increase the precision by about four bits without using a multiplier. It is shown that an improvement can be made in Taylor's technique by choosing a different constant. Also, by combining one of Arnold's techniques with a variant of Taylor's modification, precision similar to that of Taylor's design can be obtained using a simpler circuit.>

A new characterization of the interior of the convex hull of a finite point set is given. An inclusion test based on this characterization is, on average, almost linear in the number of points times the dimensionality.

We adopt the following measures of clustering based on simple edge counts in an undirected loop-free graph. Let S be a subset of the points of the graph. The compactness of S is measured by the number of edges connecting points in S to other points in S. The isolation or separation of S is measured by the negative of the number of edges connecting points in S to points not in S. The subset S is a cluster if it is compact and isolated. We study the cluster search problem: find a subset S which maximizes a linear combination of the compactness and (negative) isolation edge counts. We show that a closely related decision problem is NP-complete. We develop a pruned search tree algorithm which is much faster than complete search, especially for graphs which are derived from points embedded in a space of low dimensionality.