Graphitti focuses on performance analysis and improvements of large scale graph processing systems on various platforms. Specifically, we design, implement, tune, and analyze algorithms for graph processing on GPUs  and multi-cores, homogeneous and heterogeneous clusters, and clouds using several programming models, such as OpenCL/CUDA, MPI, Hadoop, Giraph etc.

There are five main research directions in this project.

• State-of-the-art analysis - focuses on analyzing state-of-the-art graph processing platforms.
• Data analysis - focuses on designing optimal data set structures for large graphs, and analyzing their properties.
• Cluster and Clouds - focuses on implementing solutions for large scale graph processing in clusters and clouds.
• HPC systems - focuses on implementing a Pregel-like system to make use of high performance computing resources such as BlueGene. 
• Multi-cores and GPUs - focuses on designing and tuning high-performance large scale graph processing algorithms for massively parallel architectures such as GPUs and multi-cores.

There are eight people involved in this project:

• Dr. Ana Lucia Varbanescu - PostDoc, TUDelft and VU Amsterdam
• Dr. Alexandru Iosup - Assistant Professor, TUDelft
• Yong Guo - PhD student, TUDelft
• Claudio Martella - PhD student, VU Amsterdam
• Nefeli Papapetrou Lampraki - MSc student, TUDelft 
• Marcin Biczak - MSc student, TUDelft 
• Ate Penders - MSc student, TUDelft
• Aurelian Bogdan - MSc student, VU Amsterdam