Abstract
Truncated increase decreases some portion of the power required by multipliers by just registering the most-noteworthy bits of the item. The most widely recognized way to deal with truncation incorporates physical diminishment of the halfway item network and a pay for the lessened bits by means of various equipment pay sub-circuits. Be that as it may, this outcome in settled frameworks advanced for a given application at configuration time. A novel way to deal with truncation is proposed, where a full exactness multiplier is executed, however, the dynamic segment of the fractional item network is chosen progressively at run-time. This permits a power lessening trade-off against flag corruption which can be changed at runtime. Such an engineering unites the power decrease profits by truncated multipliers and the adaptability of reconfigurable and universally useful gadgets. Proficient usage of such a multiplier is displayed in a custom computerized flag processor where the idea of programming remuneration is presented and dissected for various applications. Test results and power estimations are considered, including power estimations from both post-blend reproductions and a created IC usage. The article proposes the single and two-way pipeline truncated multiplier implemented on FPGA platform. The proposed design shows the significant improvement in the throughput of the architectures.
Keywords-component: Truncated Multiplier, SNR, Round-Off, Pipe-line Multiplier, FPGA, ASIC, Multiply Accumulate Unit (MAC)..etc
References
1) M.G. Solaz, W. Han and R. Conway , “A Flexible Low Power DSP with a Programmable Truncated Multiplier,” IEEE Transactions on Circuits and Systems , vol. 59 No. 11, pp. 2555–2568, Nov 2012.
2) N. Yoshida, E. Goto, and S. Ichikawa, “Pseudorandom rounding for truncated multipliers,” IEEE Trans. Comput., vol. 40, no. 9, pp. 1065–1067, 1991
3) M. de la Guia and R. Conway, “Data wordlength reduction in 90 nm multipliers,” in Proc. IET Irish Signals Syst. Conf. (ISSC 2009), Jun.
4) 2009, pp. 1–6 R. Conway, “Fixed-point arithmetic,” in Proc. Irish Signals Syst. Conf., 2003
5) K. Han, B. Evans, and E. E. Swartzlander, Jr., “Low-power multipliers
6) with data wordlength reduction,” in Conf. Rec. 39th Asilomar Conf.
7) Signals, Syst. Comput., 2005., Nov. 1, 2005, vol. 28, pp. 1615–1619.