INDUSTRY

MOUNTAIN VIEW, Calif. -- SGI today announced that the National Cancer Institute's Advanced Biomedical Computing Center (ABCC) in Frederick, Md., has purchased an SGI(R) Origin(R) 3800 server, three SGI(R) Origin(R) 300 servers and an SGI(R) TP9400 storage system to provide state-of-the-art computing resources to the nation's leading biomedical researchers. The new SGI(R) servers and storage system will be used to support research in such challenging areas as deciphering the human genome, understanding structure-function relations of biological molecules and developing therapeutics for numerous diseases, especially AIDS and cancer. The most powerful SGI compute servers, including a 64-processor Origin 3800 system with 64GB of memory, have been installed at the ABCC. In addition, three 8-processor Origin 300 servers, each with 8GB of RAM, were added to the ABCC's existing 40-processor Origin 300 cluster. The TP9400 storage system complements the SGI(R) Origin(R) servers with 5.6TB of disk space in a fully redundant configuration. The TP9400 system's storage is available to the servers via a 2GB Fibre Channel SAN and CXFS(TM) (clustered XFS(TM)), the industry's fastest shared filesystem. CXFS provides no-compromise data sharing, enhanced workflow and reduced costs in data-intensive environments by eliminating file duplication and the time it takes to move large files over the network. A SunFire(TM) V880 server at ABCC will also be connected to the SAN and the CXFS cluster through a CXFS for Solaris(TM) client. CXFS is a breakthrough, high-performance shared filesystem that provides heterogeneous SAN environments, allowing multiple hosts and operating systems to have direct access to physically shared disks. "The ABCC chose the SGI Origin 3800 system due to its ability to run 64 processors with 64GB of memory as a single, shared-memory system," said Dan Stevens, life and chemical sciences market manager, SGI. "A major factor in the ABCC's decision was the availability of the CXFS filesystem, which allows the Origin 3800 server to integrate the filesystem with the eight Origin 300 servers. This allows greater flexibility for load balancing, data sharing and application software maintenance within a high-performance environment." With the genomic and proteomic data explosion and the increase in data complexity, biology has become primarily a data-driven science. Therefore, high-performance disk storage, spread across several large compute nodes, is necessary to analyze all of the data and update processed resources so that other applications and servers can access the latest data to answer a user query in the next step of a pipelined environment. "With 64 processors and 64GB of memory in a single image configuration, Origin 3800 is ideally suited for handling large data sets and processing them in parallel," Stevens said. "Often memory access is a limiting factor in biological data processing and systems modeling. The simplification of software development due to the shared-memory system allows for a shorter time to solution and greater efficiency in processing for large genomic and proteomic data sets and models." Some examples of computational projects that the ABCC supports are: -- Modeling of anti-cancer drug interactions with known tumor targets -- Analysis of genomic data for chromosomal rearrangements and deletions that may lead to cancer or other diseases -- Modeling of ion channels and transmembrane receptors that are involved in heart function, cell proliferation, neurotransmitters and drug resistance in cancer -- The study of DNA-protein interactions involved in cancer and gene regulation -- Modeling of the detailed reaction mechanisms of chemical reactions involved in cell signaling and are known to be disrupted in certain forms of cancer -- Analysis of cell responses to drug challenges in the NCI tumor cell line to search for molecular properties that may serve as useful therapeutics and to understand the different interactions between cell lines About ABCC The National Cancer Institute's supercomputing facility is a fully integrated, high-performance, scientific computing resource located in Frederick, Md. Organizationally identified as the Advanced Biomedical Computing Center, the facility provides state-of-the-art computing support and technology to the scientists of the NCI, the National Institutes of Health (NIH) and extramural biomedical researchers. NIH and NCI have identified high-performance computing as an important area for biological research. The ABCC was established in 1986 by NCI to provide a high-performance computing and user-friendly environment for biomedical research. Unlike some other high-performance centers, ABCC only supports research directed toward biological problems.