Research in System Architecture and Cluster Software
Our flagship VERTEX architecture is at the forefront of research and innovation on scaling, ease of use, IO performance and hierarchical shared memory on hybrid multi-core clusters targeted toward small to very large HPC system installations (Tera FLOPs to Exa FLOPs). Among others, we have joined hands with a European oil and gas major, and system vendors Supermicro and Boston Ltd. to study the scaling and ease of use of VERTEX for applications in the petroleum and energy industry. We are working with academics at Univ. of Rochester, Univ. of Massachusetts and Purdue to extend the VERTEX platform to scalable hybrid clusters that include x86, GPUs and the IBM Cell processor.
Download VERTEX White Paper from SC10 in New Orleans.
Research in Quantitative Finance and Genome Sequencing Applications
We are investing in sophisticated quantitative finance algorithms and strategies with our partners India Institute of Quantitative Finance (IIQF) and parallelising them on CPU and GPU clusters in order to enhance the capabilities of our QF Links product. We are parallelizing several open source genome sequencing packages on CPU and GPU clusters for our Genome Links product.
Weather/Climate Applications Research
Working with researchers at Indian Institute of Technology Delhi, we are cloud enabling a few widely used weather forecasting applications such as WRF. We are also collaborating with the Monsoon Mission, a joint Indo-US govt. project including NOAA and Indian Institute of Tropical Meteorology on a pilot to move their newest weather/climate modelling application CFS into a community model. We are extending this work in a joint pilot with our global partners NVIDIA to parallelise and optimize core portions of the CFS code on GPU clusters. These research investments continue to help HPC Links in generating services revenue in these domains.