Power and Area Efficient Cascaded Effectless GDI Approximate Adder for Accelerating Multimedia Applications Using Deep Learning Model

Table 9

Comparison of clock cycles taken for computation by the convolution layers in the LWDD network.

Stage of processing	Number of clock cycles				Speed-up factor
	LWDD system based on different adders
	GDI-CMBAI	GDI-AMBAII	GDI-HSETAII	Proposed GDI-EAHSETAII	GDI-AMBAII	GDI-HSETAII	Proposed GDI-EAHSETAII
Convolution layer 1	12605	12120	10245	9771	1.04	1.23	1.29
Convolution layer 2	50416	48470	40990	39112	1.04	1.23	1.29
Convolution layer 3	25569	24587	20789	19977	1.04	1.23	1.29
Convolution layer 4	51136	46171	41560	39609	1.04	1.23	1.29
Convolution layer 5	27009	25976	21956	20885	1.04	1.23	1.29
Convolution layer 6	52016	50015	42290	40320	1.04	1.23	1.29