A NOVEL ARCHITECTURE OF LUT DESIGN OPTIMIZATION FO
no vote
Several memory based algorithms for DSP implementation have been reported in the literature, involving orthogonal transform and digital filter. Multiplication is in signal processing and ALU. The main operation of multiplier is LUT. However, we did not find any significant working memory LUT optimized multiplication. This paper presents a new design method of look-up table, in which only odd multiple storage (OMS) is used. In addition, the anti symmetric product coding (APC) method, LUT reduced by half, provides a reduction. When APC method is combined with OMS technology, the operation of two complements can be simplified, because the input address and LUT output can be converted into odd numbers, thus reducing the conventional LUT size by a quarter lut.the LUT multiplier presents word size = w = 5 and 6 bits encoded by Verilog and synthesized in Xilinx 13.4. It is found that LUT based multiplier is the same size and time complexity as 8-bit word, but higher word size, which involves less multiplication time than CSD based multiplier. For 16 bit and 32-bit word sizes, respectively, provide delay products over area savings in the corresponding CSD multiplier of 30% and 50%.