Multiply Accumulator


Important Notice:

Obsoleted in 2014.1 and not recommended for new designs.

Functionality can be implemented using RTL inference directly.

Product Description

The Multiply Accumulator IP accepts two operands, a multiplier and a multiplicand, and produces a product (A*B=Prod) that is added/subtracted to the previous adder/subtracter result (S=S+/-Prod). This product value can be loaded with assertion of Bypass (S=Prod). The Multiply Adder IP is implemented using Xtreme DSP™ slices and operates on signed or unsigned data.

Key Features and Benefits

  • Supports multiplier inputs ranging from 1 to 31 bits unsigned or 2 to 32 bits signed and an output width ranging from 1 to 79 bits unsigned or 2 to 80 bits signed 
  • Latency can be set for optimal speed or the minimal amount of pipelining allowed "Latency = 1" (accumulation register required)
  • Instantaneous Resource Estimation
  • For use with AMD CORE Generator™



Featured Documents

Default Default Title Document Type Date