Force Directed Graph Placement
Force Directed Placement: GPU Implementation 1Tripti Sharma, 2Akshita Sharma, 3Ashita Malik, 4Tushar Verma Department Of Information Technology
Maharaja Surajmal Institute Of Technology,New Delhi,India 1triptionline@yahoo.com, 2akshita.msit6@gmail.com, 3malikashita@gmail.com, 4tushar.msit27@gmail.com
Abstract - Graph layout has had important applications in many areas of computer science. When dealing with machine generated data, we often tend to see the data to have better understanding of a structural form. Many such data can be represented in the form of graphs. By laying out a graph, we can untangle information and intuitively show relations of objects. Automated graph drawing remains a difficult placement and layout problem. This problem is difficult in part due to the complexity of formulating good algorithms to draw graphs which are aesthetically pleasing for human visualization. In this paper, we have implemented a Force-Directed Placement algorithm using Python environment that will solve the graph layout problem by using an energy minimization technique. This work aims to demonstrate the performance advantage of a GPU implementation as compared to a CPU implementation. For CPU Python is the programming platform chosen and for GPU NVIDIA PyCUDA is the platform chosen. The GPU implementation was able to achieve up to a 55-60x speed-up as compared to the CPU.
Keywords-Force directed placement, PYCUDA, PYTHON
1. INTRODUCTION
Graphs are often used to encapsulate relationship between objects. Graph drawing enables visualization of such relationships. The usefulness of this relationship is dependent on whether the drawing is aesthetic. While there are no strict criteria for aesthetics of a drawing, it is generally agreed for example, that such a drawing has minimal edge crossings with vertices evenly distributed in space, and with
Maharaja Surajmal Institute Of Technology,New Delhi,India 1triptionline@yahoo.com, 2akshita.msit6@gmail.com, 3malikashita@gmail.com, 4tushar.msit27@gmail.com
Abstract - Graph layout has had important applications in many areas of computer science. When dealing with machine generated data, we often tend to see the data to have better understanding of a structural form. Many such data can be represented in the form of graphs. By laying out a graph, we can untangle information and intuitively show relations of objects. Automated graph drawing remains a difficult placement and layout problem. This problem is difficult in part due to the complexity of formulating good algorithms to draw graphs which are aesthetically pleasing for human visualization. In this paper, we have implemented a Force-Directed Placement algorithm using Python environment that will solve the graph layout problem by using an energy minimization technique. This work aims to demonstrate the performance advantage of a GPU implementation as compared to a CPU implementation. For CPU Python is the programming platform chosen and for GPU NVIDIA PyCUDA is the platform chosen. The GPU implementation was able to achieve up to a 55-60x speed-up as compared to the CPU.
Keywords-Force directed placement, PYCUDA, PYTHON
1. INTRODUCTION
Graphs are often used to encapsulate relationship between objects. Graph drawing enables visualization of such relationships. The usefulness of this relationship is dependent on whether the drawing is aesthetic. While there are no strict criteria for aesthetics of a drawing, it is generally agreed for example, that such a drawing has minimal edge crossings with vertices evenly distributed in space, and with
References: [6] P. Eades, ‘A heuristic for graph drawing’, Congressus Nutnerantiunt,42, 149–160 (1984). [8] Y. Frishman, A. Tal ‘Online Dynamic Graph Drawing’ IEEE-VGTC Symposium on Visualization (2007)