Yevgen Voronenko

Online Multiplier Block Generator

Parameter	Value	Explanation
Constants	0.2303778133088 0.7148465705529 0.6308807679298 0.0279837694168 -0.1870348117190 0.0308413818355 0.0328830116668 -0.0105974017850	Integer or floating point constants. Maximum 25 bits, 20 constants. Floating point constants will be converted into integers using the number of fractional bits given below. (?)
Fractional bits	0 (integer constants) 123 456 78 91011 121314 151617 181920 212223 2425	Constants will be scaled by 2^f, where f is the value given here. (?)
Algorithm	Hcub (our algorithm) BHM RAG-n (19 bits max)	MCM algorithm to use. RAG-n is limited to 19 bit constants. (?)
Est(z) (Hcub)	csd-cost(z) bits(z) = ceil(log2(z)) 2*bits(z) + csd-cost(z)	Auxiliary distance estimate, might affect the solution quality in Hcub, but for algorithms it is unused. (?)

Papers

Yevgen Voronenko and Markus Püschel,
Multiplierless Multiple Constant Multiplication
to appear in ACM Transactions on Algorithms
[ PDF | Figures (PNG) ]

D. R. Bull and D. H. Horrocks.
Primitive operator digital filters.
IEE Proceedings G, 138(3):401-412, 1991.
[ BIB ]

A. G. Dempster and M. D. Macleod.
Use of minimum-adder multiplier blocks in FIR digital filters.
IEEE Transactions in Circuits and Systems-II: Analog and Digital Signal Processing, 42(9):569-577, 1995.
[ BIB ]

A. G. Dempster and M. D. Macleod.
Constant integer multiplication using minimum adders.
IEE Proceedings - Circuits, Devices and Systems, 141(5):407-413, 1994.
[ BIB ]