The lab has about 2000 sq. ft. space. It composes of abundant computing faculties for medical imaging and daily operational supplies. The major components are
Medical Imaging Computational Grid (MICG)
Medical Imaging Application Software (MIAS)
High Performance Computing Technology (HPCT)
Laboratory Office and Research Space (LORS)
Copyright © 2005-2011 Laboratory of Medical Imaging High Performance Computing & Informatics; www.uiowa.edu/mihpclab/ |
Medical Imaging Computational Grid (MICG)
MICG is the medical imaging computational grid configured using grid computing technology. It is a computational facility for medical imaging. It also serves as a test-bed for medical imaging informatics system (including experimental PACS, distributed networks and communication, remote image display etc. The integrated system is designed to facilitate future distributed and centralized high performance computing (HPC) for medical image processing, data analysis, and medical imaging data mining, and other intensive computations in modeling and simulations of complex biomedical systems. It is one of three components of MiHi lab's facility.
Technical protocols: the system contains multiple distributed HPC clusters (Grid nodes) and each cluster consists of multiple computing nodes or processors. The computational jobs can be conducted and performed in sequence or in parallel cross platforms using the middle-ware for job scheduling and job monitoring facilitated by Ganglia, Sun Grid Engine (SGE), or Condor, and message passing interface (MPI) parallel library. The clusters are operated by Rocks based on Red-Hat Linux kernel. With the Rocks installation, couple useful middle-ware tools are also installed, including AREA51 for system security related services and utilities, BIO for Bioinformatics utilities, CONDOR as High throughput computing tools, GANGLIA, as a cluster monitoring system from UCB, Globus 4.0.4 (GT4) for Grid computing, Sun Java SDK and JVM for Java developers and environment, PBS, the Torque/Maui job queuing system (packaged by University of Tromso), PGI compiler (packaged by Cluster Corp), PVFS2 File System, and Sun Grid Engine as job queuing system ad installed on the clusters.
Some cluster nodes are built for production, while others are designed for experiments and test-beds. The technology of system installation can be found at HPCT.
Medical Imaging Application Software (MIAS)
MIAS lists some of medical imaging related application software installed on our clusters in MICG. The software installed are widely-used open-source software and/or popular libraries for medical imaging. These application software help lab researchers and developers to process research and clinical related medical images, with focus on developing high-performance, innovative models and methodologies, such as parallel medical imaging reconstruction, segregation, registration, enhancement, etc. Some of software toolkits and programming techniques also help to visualize image data such as VDK, iTK, MITK, etc. to assist medical data analysis for research or clinic diagnosis. Potential functions can be integrated together to develop a enterprise solutions for translation al science award and/or clinical trails, and for public uses.
Laboratory Office and Research Space (LORS)
MIHPC Lab is located at University of Iowa's Westlawn. It has about 1500 sq. ft , including offices, visualization workstations, hardware equipments, computing rooms, meeting room, technical library etc. All the lab clusters built by our professionals are located in the computing machine room in the lab. The spotlights of the lab can be virtually visited by clicking the following lab pictures..