3D Data Denoising via Nonlocal Means Filter by Using Parallel GPU Strategies

Table 1

L1-prefer switch influence on execution times for the partial and full unrolling algorithm with (128, 1, 1) block size configuration and 3D datasets of normally distributed random numbers.

	Partial unrolling algorithm				Full unrolling algorithm
	Single GPU		Multi-GPU		Single GPU		Multi-GPU
	L1	no L1	L1	no L1	L1	no L1	L1	no L1
64^3	0.77	0.83	0.38	0.42	0.58	0.58	0.23	0.23
128^2×64	2.25	2.3	1.12	1.15	2.29	2.29	0.85	0.85
128^3	4.5	4.59	2.25	2.3	4.38	4.38	1.9	1.9
256^2×128	16.24	16.54	8.12	8.25	17.51	17.51	7.58	7.58
256^3	32.5	33.58	16.24	16.54	34.22	34.24	15.94	15.95
512^2×128	63.55	66.42	31.67	32.52	70.01	70	30.31	30.32
512^2×256	128.19	138.06	63.57	66.59	136.89	136.95	63.75	63.77
512^3	260.93	288.84	128.63	138.2	270.61	270.75	130.78	130.67