Computational aerodynamic and aeroacoustic analysis of a quiet supersonic business jet, focussing on the inlet and nozzle aerodynamics, and acoustics of the fan and nozzle.

Computational predictions of turbulent heat transfers in high speed boundary layer flows.

High-altitude plume-atmosphere interactions produces complex three-dimensional chemically reacting flows.

The group led by Prof. Merkle is involved in Multi-scale and Multi-physics Computational fluid dynamics (CFD). The group is working on several projects. Please visit our website for detailed information. The figure shows the air flow around the blades of a jet engine fan.

Our groups research focus is in low gravity and MEMS 3-D real-world capillary fluids problems. We use the freely available powerful "Surface Evolver" code for the computational simulations.

The spacecraft mission design and analysis research with very large data sets and manifold structures includes 1. optimal design and analysis of spacecraft missions including trajectory design, 2) missions involving spcae-based interferometry: including searches for exosolar planets, and 3) identification of black holes, and spectral characterization of distant stars.

Our research focuses on - Protein function beyond BLAST search, Protein 3D structure prediction, fast protein 3D structure search and protein-protein docking.

The pharmaceutical informatics research thrust area in our group aims to establish new mathematical, informatics and computational foundations to broaden the technological base for harvesting the fruits of genomics. For more information read this essay on pharmaceutical informatics - http://meweb.ecn.purdue.edu/~kim55/Pharma-essay.shtml

Our research aim is to develop novel computational concepts to identify feasible binding modes of protein-bound molecules and to compute their binding affinity. Our ongoing concept development is focused on the aspects of protein flexibility, solvation, and entropy.

Fastest N-body solvers, long-time integrators, absolute free energy differences, transition pathways

The research focuses on the development of cryo-EM technique to push for near atomic resolution and high throughput single particle 3-D reconstructions and to eventually transform this technique into a routine tool for functional studies of biological systems.To achieve these goals, our research involves development of new image processing algorithms, high performance computing, data collection automation and reliable sample preparation.

Changes in extreme hot events in the 21st century show an increase of of 100 to 560% in occurrence of the hottest temperatures.

High resolution, coupled models for global quantification of carbon dioxide emissions.

The Ecosystems and Biogeochemical Dynamics group is studying the interactions among atmosphere, biosphere, and human dimension in the context of climate change, chemical element cycles, and policy-making. To study these interactions, the data of biosphere and atmosphere obtained from field and in-situ measurements and satellite observations are fused with the numerical models of the ecosystems and biogeochemistry and the atmosphere.

Seismic receiver function array imaging of the subduction zone in the Pacific Northwest beneath Oregon. Figure shows the ray migration imaging using a single teleseismic event.

The goal of this project is to develop suitable tools to help understand mass transfer processes during air advective air flux, in order to aid in the design and operation of remediation systems that involve advective air flux such air sparging and Soil vapor Extraction (AS/SVE).

The focus areas for CSCAPES are load balancing in parallel computation, automatic differentiation, advanced methods for sparse matrix computations, and runtime data and iteration reordering to improve performance in irregular computation.. This will accelerate the development and deployment of fundamental enabling technologies in high performance computing, by creating algorithms and software tools for key combinatorial problems in scientific computing at the petascale.

The Compact Muon Solenoid Detector (CMS) of the Large Hadron Collider (under construction at CERN) will enable new physics in the area of high energy physics. The CMS center at Purdue is a tier 2 center providing storage for the large amounts of data that will be generated from these experiments and also provide other services (including computing) to other researchers.

SPIKE is a robust hybrid parallel banded linear system solver that can be used as a direct solver or pre-conditioner for outer iterative scheme. The banded system could be dense or sparse. We have seen performance better than Scalapack package in different architectures.

Over a number of years our research group has been working on auto-tuning compilers with the goal of making programs run optimally on wide array of architectures by optimally searching a huge space of optimization options, dynamically plug-in new code while ensuring the evolving program improves.

The goals of this project is to exploit chip multi-processor (CMP) cores by exploiting the tradeoff of latency and capacity. With more CMP cores we need both large capacity and fast access from on-chip caches.

Our research group looks at addressing computation and communication issues in modern HPC architectures and interconnection networks.

USFS is a peer-to-peer (p2p) enhancement for the widely-used Network File System (NFS). USFS harvests redundant storage space on cluster nodes and user desktops to provide a reliable, shared file system that acts as a large storage with normal NFS semantics

We are looking at reliability issues on grid. Our solution to these issues uses techniques such as predicting failure (using Semi-Markov model and Neural network), making scheduling decisions, perform runtime monitoring and Incremental failure prediction, migrate processes from failure-prone hosts using checkpoints

The objective of the PRISM project is to significantly accelerate the integration of MEMS technologies into stockpile monitoring and weapons systems through the use of predictive, validated science and petascale computing. We seek to understand, control, and improve the long-term reliability and survivability of MEMS by using multiscale multiphysics simulation, from atoms to micro-devices, to address fundamental failure mechanisms. The central focus is on a single class of contacting radio-frequency (RF) metal-dielectric capacitative MEMS switches that will impact development of civilian and military MEMS devices of the future.
Web site: http://www.purdue.edu/discoverypark/prism/about.shtml

Our research group uses multi-scale modeling to predict behavior of materials from first principles and understand the fundamental mechanisms that govern materials behavior.

The nanoelectronic modeling of multi-million atom electronic structure calculations using state of the art NEMO 3-D application is carried out in our group. It can calculate eigenstates in (almost) arbitrarily shaped semiconductor structures in the typical column IV and III-V materials. For more details please visit our website.

The center combines both advanced integrated computational tools, and state-of-the art experimental devices. The computational tools are multipurpose multidimensional computational tools that integrate various fields of science and engineering such as heat transfer, thermal hydraulics, magneto-hydrodynamics, atomic and plasma physics, photon transport, and material erosion lifetime. On the experimental side, initially three primary components and devices will make up the verification tools. One is a high-intense charged-particle/material interaction experiments with unique in-situ metrology for mixed materials surface physics studies. The second is high-intensity lasers for intense power deposition simulation experiments and the third, a high-power Z-pinch plasma source for plasma-material interaction and modifications studies to create super materials and alloys for various applications including energy, biomedical, nanotechnology, advanced lithography, directed energy and national security, and nuclear and high energy physics.

The goal of this project is to address challenging questions related to performance of real application benchmarks such as SPEC MPI 2007 on different high performance computing platforms such as Cray-XT4, IBM Blue Gene, SGI Altix, IBM JS21 linux cluster etc., We use these benchmarks for benchmarking and performance characterization of new architectures. We also maintain a rank list based on such real application benchmarks.