Skip to main content

Table 2 Average frame rate (in frames per second) of the execution of marching cubes algorithm in CPU and GPU, with different data structures, comparing against the brute force version

From: Marching cubes technique for volumetric visualization accelerated with graphics processing units

Volume name

Brute force

k-d tree

Interval tree

Quadtree

Octree

CPU

   Fuel

83.3

105.6

114.4

108.0

98.3

   Hydrogen atom

17.5

23.4

25.6

24.0

19.2

   Angiography

7.5

12.9

14.1

13.6

10.7

   Engine

4.6

5.2

5.7

5.5

4.8

   Lobster

3.2

3.7

4.1

3.9

3.6

   Head

1.8

2.0

2.3

2.1

1.9

   Bonsai

1.2

1.3

1.6

1.4

1.3

   Knee

1.1

1.2

1.4

1.3

1.1

   Sheep heart

0.8

0.9

1.0

0.9

0.8

   Aneurism

0.6

0.9

1.3

1.2

0.7

GPU

   Fuel

195.4 (2.3)

220.7 (2.1)

246.4 (2.2)

229.9 (2.1)

205.7 (2.1)

   Hydrogen Atom

77.0 (4.4)

142.8 (6.1)

165.0 (6.4)

149.5 (6.2)

96.5 (5.0)

   Angiography

24.7 (3.3)

109.6 (8.5)

133.7 (9.5)

89.2 (6.6)

58.7 (5.5)

   Engine

26.6 (5.8)

61.0 (11.7)

81.1 (14.2)

56.8 (10.3)

38.4 (8.0)

   Lobster

30.1 (9.4)

42.8 (11.6)

58.5 (14.3)

29.8 (7.6)

27.4 (7.6)

   Head

17.9 (9.9)

22.7 (11.4)

35.8 (15.6)

28.9 (13.8)

20.9 (11.0)

   Bonsai

10.7 (8.9)

13.2 (10.2)

25.9 (16.2)

18.7 (13.4)

11.5 (8.8)

   Knee

8.2 (7.4)

9.7 (8.1)

22.4 (16.0)

10.7 (8.2)

9.1 (8.2)

   Sheep Heart

6.3 (7.9)

7.4 (8.2)

16.3 (16.3)

8.9 (9.9)

6.6 (8.3)

   Aneurism

2.1 (3.5)

10.2 (11.3)

23.7 (18.2)

11.9 (9.9)

9.5 (13.5)

  1. Values between parenthesis correspond to the acceleration factor related to their respective data structure versions of marching cubes in CPU for each volume dataset