Table 2 Ln CORDIC average error analysis and comparison based on input range and number of iterations.

From: Digital design of a spatial-pow-STDP learning block with high accuracy utilizing pow CORDIC for large-scale image classifier spatiotemporal SNN

CORDICs for ln

input range

n_ln = 2

n_ln = 4

n_ln = 8

n_ln = 10

n_ln = 12

conventional (Eq. 19)

\([0.1068 , 9.3595]\)

\(1.72\times {10}^{0}\)

\(5.61\times {10}^{-1}\)

\(2.71\times {10}^{-1}\)

\(2.16\times {10}^{-1}\)

\(2.58\times {10}^{-1}\)

Luo’s CORDIC40

\([0.107 , 9.352]\)

\(3.8023\times {10}^{-4}\)

Mopuri's CORDIC44

\([9.6358\times {10}^{-7} , 1.0378\times {10}^{6}]\)

\(4.7\times {10}^{-3}\)

\(3.0735\times {10}^{-4}\)

\(1.923\times {10}^{-5}\)

Chen's CORDIC43

\([6.3471\times {10}^{-8} , 6.8351\times {10}^{4}]\)

\(6.35\times {10}^{-5}\)

Ln CORDIC (Proposed)

\(\left(0 , {10}^{9}\right] (0 ,+\infty )\)

\(1.21\times {10}^{-4}\)

\(1.61\times {10}^{-6}\)

\(3.81\times {10}^{-10}\)

\(5.93\times {10}^{-12}\)

\(9.37\times {10}^{-14}\)