Research on Parallel Simulation Method of Underwater Wireless Optical Channel

Fig. 10. Impulse response for different water types in computer environment 1. (a) Computing platform is Matlab; (b) computing platform is Open MP; (c) computing platform is CUDA

Download full size

Fig. 11. Impulse response for different water types in computer environment 2. (a) Computing platform is Matlab; (b) computing platform is Open MP; (c) computing platform is CUDA

Download full size

Table 1. Optical properties of different water types
View table
Table 1. Optical properties of different water types
Water type $μ_{a} / m^{- 1}$ $μ_{s} / m^{- 1}$ $μ_{t} / m^{- 1}$ $ω_{0}$
Clear ocean（Ⅰ） 0.114 0.037 0.151 0.245
Coastal（Ⅱ） 0.179 0.291 0.398 0.731
Harbor（Ⅲ） 0.366 1.824 2.190 0.833

Table 2. Two computer environments
View table
Table 2. Two computer environments
Computer environment 1 Computer environment 2
GeForce MX 150 GeForce GTX 1050 TI
CUDA core 384 768
Clock rate /MHz 936 1366
Global memory /Mb 6098 12252
CPU i7-8550U E5-2670
CPU core 4 8

Table 3. Test parameters and test results in different waters, platforms and computing environmental conditions

View table

Table 3. Test parameters and test results in different waters, platforms and computing environmental conditions

		Simulation parameter			Result
Computer environment	Computing platform	Photon number	Distance /m	Water type	MSE	$R^{2}$
1	Matlab	$3 \times 10^{6}$	30	Ⅰ	0.012760	0.9616
		$3 \times 10^{6}$	30	Ⅱ	0.069580	0.9156
		$3 \times 10^{6}$	30	Ⅲ	0.052060	0.9594
	Open MP	$3 \times 10^{6}$	30	Ⅰ	0.005969	0.9906
		$3 \times 10^{6}$	30	Ⅱ	0.060130	0.9137
		$3 \times 10^{6}$	30	Ⅲ	0.041380	0.9832
	CUDA	$3 \times 10^{6}$	30	Ⅰ	0.004200	0.9962
		$3 \times 10^{6}$	30	Ⅱ	0.061600	0.9295
		$3 \times 10^{6}$	30	Ⅲ	0.059200	0.9567
2	Matlab	$3 \times 10^{6}$	30	Ⅰ	0.010900	0.9728
		$3 \times 10^{6}$	30	Ⅱ	0.059800	0.9280
		$3 \times 10^{6}$	30	Ⅲ	0.040100	0.9846
	Open MP	$3 \times 10^{6}$	30	Ⅰ	0.005900	0.9914
		$3 \times 10^{6}$	30	Ⅱ	0.062100	0.9163
		$3 \times 10^{6}$	30	Ⅲ	0.053900	0.9563
	CUDA	$3 \times 10^{6}$	30	Ⅰ	0.003300	0.9976
		$3 \times 10^{6}$	30	Ⅱ	0.059200	0.9842
		$3 \times 10^{6}$	30	Ⅲ	0.052500	0.9604

Table 4. Acceleration effect in computing environment 1

View table

Table 4. Acceleration effect in computing environment 1

Number of photons	Distance /m	Speedup of M to O	Speedup of M to C	Speedup of O to C
$1 \times 10^{5}$	20	16.21	101.26	6.25
$1 \times 10^{5}$	35	16.23	75.33	4.64
$2 \times 10^{5}$	20	33.92	116.98	3.45
$2 \times 10^{5}$	35	58.83	189.31	3.22
$3 \times 10^{5}$	20	52.75	132.38	2.51
$3 \times 10^{5}$	35	40.53	121.85	3.01

Table 5. Acceleration effect in computing environment 2

View table

Table 5. Acceleration effect in computing environment 2

Number of photons	Distance /m	Speedup of M to O	Speedup of M to C	Speedup of O to C
$1 \times 10^{5}$	20	32.04	243.20	7.59
$1 \times 10^{5}$	35	33.66	234.44	6.97
$2 \times 10^{5}$	20	54.59	351.56	6.44
$2 \times 10^{5}$	35	44.24	351.17	7.94
$3 \times 10^{5}$	20	93.05	210.86	2.27
$3 \times 10^{5}$	35	76.93	207.76	2.70

Table 6. Acceleration effect of different photon number

View table

Table 6. Acceleration effect of different photon number

Number of photons	Type of sea water	Distance /m	Running time of Open MP /s	Running time of CUDA /s	Speedup of C to O
$1 \times 10^{5}$	Ocean water（I）	50	5.074	0.477	10.64
$3 \times 10^{5}$		50	14.485	3.786	3.83
$5 \times 10^{5}$		50	21.491	7.844	2.74
$1 \times 10^{6}$		50	43.959	10.877	4.04

Table 7. Acceleration effect of different distances

View table

Table 7. Acceleration effect of different distances

Number of photons	Type of sea water	Distance /m	Running time of Open MP /s	Running time of CUDA /s	Speedup of C to O
$3 \times 10^{5}$	Ocean water （I）	10	12.231	3.776	3.24
$3 \times 10^{5}$		30	12.755	3.781	3.37
$3 \times 10^{5}$		50	13.683	3.988	3.43
$3 \times 10^{5}$		100	15.764	5.872	2.68

Table 8. Acceleration effect of different water types

View table

Table 8. Acceleration effect of different water types

Number of photons	Type of sea water	Distance /m	Running time of open MP /s	Running time of CUDA /s	Speedup of C to O
$3 \times 10^{5}$	Ocean water（Ⅰ）	30	7.287	3.705	1.97
$3 \times 10^{5}$	Coastal（Ⅱ）	30	8.005	3.903	2.05
$3 \times 10^{5}$	Harbor（Ⅲ）	30	10.003	4.002	2.50

Table 9. Acceleration effect of different optimization methods

View table

Table 9. Acceleration effect of different optimization methods

				Computer environment 2		Computer environment 1
	Photon number	Distance /m	Water type	Running time of CUDA /s	Running time of Open MP /s	Running time of CUDA /s	Running time of Open MP /s
Non-optimized	$3 \times 10^{6}$	30	Ⅱ	831	1051	1059	1347
Optimized	$3 \times 10^{6}$	30	Ⅱ	422	659	562	723
Non-optimized	$3 \times 10^{6}$	30	Ⅲ	917	1194	1084	1471
Optimized	$3 \times 10^{6}$	30	Ⅲ	562	714	633	830

Tools

Get Citation

Copy Citation Text

Jianlei Zhang, Linlin Kou, Jie Wang, Xuechen Liu, Yi Yang, Fengtao He. Research on Parallel Simulation Method of Underwater Wireless Optical Channel[J]. Laser & Optoelectronics Progress, 2022, 59(19): 1901001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites