Current Status

The installation plan for the fourth quarter is available at http://www.cs.purdue.edu/homes/jtk/intel/estatus.htm. The plan calls for the acquisition of 21 dual Pentium II systems and one quad Pentium Pro system. Of these, two dual processor Pentium II systems and one Quad Pentium Pro system have been delivered. The acquisition plan also calls for a Proshare conference manager and a Proshare video system. These will be used to test multimedia applications of clustered platforms. The machines acquired in the fourth quarter of 1997 will be used primarily as front ends to the compute server which will be developed as a cluster of Pentium II machines. These machines will be acquired in phases starting the first quarter of 1998.

Planned Activities

Hierarchical Methods for Particle Dynamics and Dense Linear System Solvers [Grama]

This project aims to develop tools and techniques for using clustered platforms for hierarchical n-body computations and their applications in solving dense linear systems. We are currently developing latency tolerant algorithms capable of delivering high processor performance and parallel efficiency. These formulations are based on function shipping paradigms as opposed to the traditional data shipping techniques.

An important application of hierarchical methods is in the solution of dense linear systems in computational electromagnetics. In our previous research, we have developed hierarchical dense solvers for tightly coupled parallel computers such as the Cray T3D. We will adapt these formulations to clustered platforms. We will also use these to solve hybrid sparse-dense system arising from a finite element - boundary element modeling of complex inhomogeneous domains.

Application Environments for Large-Scale Applications [Rice/Houstis]

This project aims to develop problem solving environments (PSEs) technology and software systems as applied to gas turbine dynamics. These engines are very complex (with ~30,000 parts) and have extreme operating conditions (with 10-50,000 RPM, 1000-1500F temperatures, pressures of 20-50 atmospheres, and 5-10G loads). The important physical phenomena take place on scales from 10-100 microns to meters. The primary goal of this research is to advance the state-of-the-art in very complex scientific simulations and their validation. The application cluster will be used to program a variety of computationally intensive components of this PSE. It will also be used to implement distributed PSE servers.

Performance Modeling of Large-Scale Computing Systems [Rice/Houstis]

The POEMS project will create and demonstrate a capability for prediction of the end-to-end performance of parallel/distributed implementations of large scale adaptive applications. The POEMS modeling capability will span applications, operating systems including parallel I/O, and architecture. Research will focus on the areas where there is little conventional wisdom available in the execution behaviors of adaptive algorithms on multi-level memory hierarchies and in parallel I/O operations. Prediction models developed as a part of this project will be validated on the clustered environment.

High Performance Algorithms in Linear Algebra [Sameh]

This project focuses on development of latency tolerant parallel algorithms for a variety of problems in linear algebra. In particular, it aims to develop solvers and preconditioners for a variety of sparse linear systems of equations and eigenvalue problems arising in simulation and design. Solvers based on the multilevel method, row-projection techniques, balanced methods, and hybrid sparse - hierarchical techniques are being investigated. These will be tested on applications ranging from computational fluid dynamics to computational electromagnetics.

Methodologies and Software Tools for Heterogeneous Environments [Rego]

The emphasis of this project is on methodologies and software tools for seamless, secure, scalable and fault-tolerant concurrent computing on heterogeneous networked computing platforms. Central to this effort is the notion of domain-oriented software support, with diverse application domains including particle-physics, combinatorial optimization and manufacturing systems. The parallel simulation of stochastic and time-stepped systems is a major component of this effort. The environments we target are heterogeneous networked platforms, including large scale, geographically distributed, collections of computers ranging from parallel processors and servers/supercomputers to (clusters of) workstations and specialized computing engines. We intend to develop enabling technologies and prototype frameworks for collaborative high-performance distributed computing and simulation that may be adapted and enhanced to deploy scalable and portable systems. We build upon our combined expertise in heterogeneous distributed processing and multithreaded fail-safe computing to provide enabling prototyping technologies for collaborative interactive systems.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

Current Status

(not available)

Planned Activities

(not available)

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

Workstations and network hubs are yet to be received. This has precluded the porting of our network computing software as well as application codes to the NT servers.

Planned Activities

We plan to port PUNCH software and applications codes during Spring and Summer of 1998 and possibly use the systems for research and education activities in Fall 1998.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	3	3	0	0
High-end Workstation	3	3	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

Equipment will be used for two purposes: (1) to meet computational demands of research and classes that will start using Intel/NT resources and (2) to further develop and experiment with PUNCH.

Current Status

(not available)

Planned Activities

(not available)

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

The equipment has been installed in the labs of Profs. Siegel and Marinescu.

Planned Activities

The focus of our research is scheduling in heterogeneous environments. We are currently examining the issues of software development in the Windows NT environment and will be porting our existing software in the coming year.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	0	6	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

Current Status

The Network Simulation and Optimization Laboratory has been established with the initial equipment deliveries. The current equipment has been configured into multiple network configurations for the purpose of generating network traffic datasets for analysis. Various network monitoring systems and network simulation tools have been secured and installed on the various systems in the laboratory. The equipment is currently being used by members of an existing enterprise network management course and by three students performing independent studies in network optimization. These initial projects are focusing on implementation of the software and data collection.

Planned Activities

The next phase of the project is to benchmark the simulation packages. Throughout the course of the spring, additional test network configurations will be built and analyzed to provide control data to compare the results of simulation runs against. The knowledge gained in these exercises will be used in the development of course materials in preparation for an initial formal course offering in Network Simulation in spring 1999.

Industry interest in the laboratory continues to escalate. A major pharmaceutical corporation has expressed interest in funding additional graduate students to expand the project to produce quality network analysts to troubleshoot and optimize their international enterprise voice and data networks.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	1	0	0
High-end Workstation	8	0	0	0
Low-end Server	0	6	0	0
High-end Server	0	0	0	0

These additional machines and Intel network equipment will allow for larger test networks to be built and analyzed, expanding the practical knowledge base and aiding course development efforts.

Current Status

Four workstations were received in December of 1997 and installed in the Robot Vision Lab, the Intelligent Robotics and Automation Lab, and the Articulated Motion Lab.

Planned Activities

We are currently working on converting our VxWorks/VME-based robotics and automation research and teaching laboratories to Windows/Intel-based systems. Since most of our software is written "C", most of this effort involves writing device drivers in Windows NT. Once the individual laboratories are converted to this common platform, we will focus on the integration of their unique capabilities. This will require sharing various modes of information, e.g., video, graphics, and process models, across the network to optimize the assembly operations required in advanced manufacturing.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

The T4E 2000 facility is a core research facility for networked high-performance computation in Biomedical Engineering as a part of the new interdisciplinary graduate program in Biomedical Engineering whose first class entered in September 1997. Two key research directions requiring such a facility are networked real-time multi-modal medical instrumentation and 3D applied physics calculations. The real-time instrumentation capability is not yet operational (see our plans for 1998). Therefore, we describe a typical applied physics problem that has been solved using Intel PCs.

Acoustics is a major emphasis in Biomedical Engineering. Under the guidance of Professor G. R. Wodicka, Dr. Martin Kompis and Yuichi Motai have developed novel experimental and computational tools for reconstructing the natural acoustic field within the lung from microphones placed on the chest surface. Note that the reconstruction is based on the natural sounds always generated by the flow of air in the lung and not on reflected or transmitted sound from an external artificial source such as is used in traditional medical ultrasound examinations. Therefore the technique is sensitive to the functioning of the lung rather than, for example, to only the anatomy of the lung and is furthermore absolutely safe. An example reconstruction for an adolescent male patient is shown in the following figure and is based on 16 chest surface microphones recording expiratory flow data at frequencies above 600Hz. The image, which is a posterior view, shows an abnormal dark region in the left lower lung lobe. This finding exactly matches the patient's diagnosis, which is blastomycosis (a fungal infection) of the left lower lung lobe, and provides a novel source of information on the extent of the disease. With continued development, this technique promises to be a novel method for making spatially-dependent measurements of lung function both in a diagnostic and in a monitoring mode.

Planned Activities

The major system development goal for 1998 is to develop the real-time multi-modal medical instrumentation network. The resources needed for hardware enhancements (A/D boards) and software enhancements (LabView) are all in place. This facility will dramatically improve the quality of existing measurements (e.g., cardiac electro-physiology measurements). Furthermore, it will allow novel integration of multiple types of measurements and the use of measurements in closed-loop control systems both of which have not been possible in the past due to inadequate computational power.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	0	5	2	2
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

The key 1998 goal for Project B.05 is to implement the real-time multi-modal medical instrumentation network. The only missing piece is high-end workstations that will host the data acquisition, processing, and synthesis of control outputs since the server has arrived, the A/D-D/A boards are on order, and the software is on-site.

Current Status

The supporting geometry and scene library middle layers of the research codes of the projects Shastra (Bajaj), EREP (Hoffmann), Artesano (Peters), HIPAIR (Sacks) have essentially been ported to the NT environment. Oglesby's group has successfully ported the WXP weather processor and visualization package to the NT environment. Turek's group used the NT environment extensively for projects related to veterinary science -- please see the web site for a list of 12 subprojects. The teaching lab was used for the undergraduate graphics course using Microsoft Visual J++ as implementation language.

More information is available in a summary presentation.

Planned Activities

The efforts for 1998 center around user interfaces and distributed processing. For interfaces, the general strategy is to build GUI interfaces using Java and VRML front ends. Distributed processing seeks to extends the environments to work on both LAN and WAN clusters of Intel platforms.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	2
High-end Workstation	2
Low-end Server	0
High-end Server	0

Current Status

The project is, unfortunately, not progressing as well as planned. In part, this is due to the lack of publicly available Java-based biocomputing programs -- we were anticipating these to be available by this time. Another factor is a lack of personnel time at Purdue and a consequent slow "ramping up" process. We do not at this time have an adequate (e.g., non-alpha test) suite of programs for our end users.

Planned Activities

During the next year we anticipate being able to devote more personnel time to the project. We also expect that the world-wide biocomputing programming effort will be more productive as programmers become more Java-aware. Because we are not currently exceeding the capabilities of our current Intel grant computer we do not expect to need further equipment in the near term. We do anticipate the need for another server in the long term, either late 1998 or 1999.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	0	0	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	1

Current Status

Synthetic Environments for Analysis and Simulation (SEAS) emulate military wargaming principles in the business and economics settings. It is a distributed, interactive simulation environment where hundreds of players can play against any combination of intelligent agents (virtual), people (live), and people helped by intelligent agents (constructive). In this synthetic environment faculty, students, and business executives can simulate decisions at the strategic, tactical, or operational levels for any firm, in any industry, in any economy. SEAS employs several emerging technologies such as intelligent agent programming, advanced 3-D visualization, net-casting, network computing, and Java3D to provide a collaborative teaching and research infrastructure. On November 21, we staged another major business wargame exercise at the Krannert School. During this exercise we simulated the Telecommunications Industry’s future, in which, approximately 180 MBA students and over 20 industry and government representatives participated. Data collected from such exercises is helping us develop "computational models of human behavior." Our next major exercise is scheduled for April 3, 1998.

Planned Activities

In the next year we plan to provide high fidelity interface to SEAS by developing a virtual gameboard (VGB). VGB will allow users to enter, manipulate, and travel through time in a computer generated, interactive, three dimensional virtual world. A virtual world is a computer generated model of reality displayed as a 3-D illusion on a computer screen using mathematical techniques such as wire framing; rendering; texture, reflection, and environment mapping; ray tracing; and animation. SEAS high fidelity interface will allow students to solve problems by seeing and experiencing information in different ways, viewing complex relationships spatially, and navigating inside synthetic markets. Students will be able to visualize continually changing financial data in "real-time" to spot trends in currency and price movements in multiple markets. A player acting as a portfolio manager or a business analyst in the game can create manipulatable images to search for relationships and patterns in large quantities of data such as actuarial tables, demographics, sales figures, and financial data.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	4	2	2	4
Low-end Server	0	0	0	0
High-end Server	0	1	0	1

Current Status

A prototype adaptive finite element model of grain drying within bins has been developed. The prototype shows excellent potential and requires further validation before it is scaled. A model to estimate preferential movement of pollutants within the soil has been developed and preliminary testing at laboratory scales completed. The high-end NT server received on this project has been setup and spatial data to support contaminant transport modeling are being developed and made WWW-accessible. An example of the spatial data for areas such as the entire state of Indiana can be found at http://abent0.ecn.purdue.edu/state/state.html. An approach for developing WWW-based expert system based tools for management of machinery systems has been developed and preliminary testing completed.

Planned Activities

The adaptive finite element model of grain drying within bins will be tested and scaled thus requiring significant computational capabilities. The preferential movement of contaminants model will be scaled to field areas as a next step. Such scaling will require significant computational resources. If success at this level, the model will later be scaled to watershed sized areas. Additional natural resources spatial data will be made WWW-accessible in support of field modeling of contaminant transport efforts. Providing data such as that at http://abent0.ecn.purdue.edu/state/state.html is computationally and data intensive. Additional testing of the ability of the NT server to handle large numbers of requests of this nature will be explored.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	3	0	4	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

The prototype models will be scaled and tested requiring significant computational capabilities.

Current Status

Our research to improve food manufacturing using high performance modeling and simulation began immediately on receipt of the 97Q3 shipment of Intel T4E equipment. To date, projects have been initiated in four distinct areas: (1) Manufacturing systems optimization using Artificial Neural Network (ANN) modeling techniques; (2) Integration of expert systems and classical methods to improve the dynamic performance in unit food manufacturing operations through multivariable adaptive control; (3) Enhancement of food safety in minimally processed foods through the development of statistical multivariate predictive growth models of pathogenic micro-organisms; and (4) Development of effective, efficient and robust Human Machine Interface (HMI) systems for obtaining plant floor data through intranet and internet access. These projects will take an additional 1-2 years to complete.

Planned Activities

We plan to begin research in additional areas upon receipt of 98Q1 and 98Q2 equipment. These research areas include: (1) The development of mathematical models for turbulent flow, two-phase flow and computer simulation of heat/equilibration/hold and cool stages of aseptically processed food systems. This research will use the Intel equipment to determine and model the thermal and rheological properties of non-Newtonian fluid foods with and without particulates in laminar and turbulent flow conditions; (2) The integration of software used to plan and implement manufacturing schedules with DCS and instrumentation fieldbus systems on the plant floor; (3) The development of an automation validation protocol for food manufacturing that considers redundant and fault-tolerant systems for hardware and software; and (4) The development of a multimedia system to educate students in food process automation. Finally, these seemingly disparate projects will be integrated to develop plant-wide models for simulation and control in our research, education and extension outreach programs.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	4	4	0	0
High-end Workstation	2	2	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

The new equipment will be used to begin the next phase of our research and facilitate workshops that will present the results of our research to the food industry.

Current Status

The NAPRA (National Agricultural Pesticide Risk Analysis) WWW-based water quality modeling system has been extended to consider nutrient losses. The NAPRA WWW-based system is being tested on 16 watersheds to determine its ability to estimate the vulnerability of watersheds to pesticide and nutrient losses. Additional modifications are also being made to the code prior to scaling its application to larger areas.

Efforts have been completed to determine how to automate parameterization of the SWAT hydrologic/water quality tool (model, GIS, databases). The modeling system will be applied to additional basin areas within the US.

Water quality routines are being developed and added to the urban runoff model LTHIA. These routines will be tested using several watersheds. The new model will be scaled for application to other areas following testing.

Planned Activities

In the coming year, the modeling systems will be evaluated initially under a range of conditions. Spatial and other supporting databases will be created to scale these systems to significantly larger areas. Several of these modeling systems have limited WWW access at the present time. Additional access to these tools will be provided via the WWW which will require significant computational resources.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	0	6	6	0
Low-end Server	0	0	0	0
High-end Server	0	1	0	0

The equipment requested will be required to scale the modeling systems for application to larger areas. The models are computationally intensive and will require significant computational power.

Current Status

Delivery of a high-end server was taken in September (97Q3), and of a low-end workstation in December (97Q4). In spite of early (and to some extent, current) delays in installation and networking, the work done in this project has progressed steadily. This has primarily been in porting and installing a variety of general-purpose software (VISUAL FORTRAN and VISUAL C++ compilers), visualization software (PV-WAVE for the Windows platform), and special-purpose software (including i) FIDAP, a finite-element package for computational fluid dynamics, ii) GEOEAS, a geo-statistical package for assessing environmental exposure, iii) NEUROSOLUTIONS, a neural network simulation using MS EXCEL as a user interface). In preparation for use of FIDAP, D. Lyn and E. Blatchley participated in training sessions in October and November at the FIDAP office in Evanston, Illinois. Students with  heavy computational requirements for their research are being encouraged to migrate to the Intel workstations, and to become familiar with the available software.

Planned Activities

In the immediate short-term, there are plans to install additional general-purpose software, such as IMSL math libraries that can be used with VISUAL FORTRAN, as well as additional special-purpose software geared to environmental/hydraulic applications, such as DYRESM, an unsteady model of lake water quality. In the longer-term, research projects, including numerical simulations of flow in disinfection channels, scour processes in rivers, flows with surfactants, and groundwater contaminant transport, will begin to exploit more fully the Intel equipment in the next year. Although initially the focus will be on research, as more experience is gathered with the available software, the use of the equipment may be broadened to include formal classes or lectures.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	1	1	0	0
High-end Workstation	1	0	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	2	2	2	2
High-end Workstation	0	0	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

These smaller machines will be used with the server and the controller to test distributed computing software

Current Status

All seven computers, including the Dual Pentium were received in the second week of December. They have since been assembled by our technical assistance staff and are now individually connected to the campus' Engineering Computer Network (ECN). We are in the process of ordering Ethernet cards for the computers so that the single processor machines may be connected to the campus network through the Dual Pentium, which will act as a local server. The software on these machines is currently being installed and is expected to be completed very soon.

Planned Activities

Planned Activities for the current year will be focussed on modifying and transporting codes currently written for UNIX based workstations to the Windows NT environment, and studying the performance of these machines vis a vis that of UNIX workstations on which these codes are currently run.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	2	2	2	2
High-end Workstation	0	0	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

Current Status

The first two Intel workstations were been received in January and installed in the Turbomachinery Fluid Dynamics Laboratory. One workstation is currently being used to reduce large quantities of velocity and pressure data from multistage compressor wake interaction measurements. The second is being used with MATLAB to analyze reduce pressure and temperature sensitive paint images made on turbomachine blades in the Large Scale Annular Cascade Facility.

Planned Activities

Dantech Measurement Technology is currently contracted for the image acquisition equipment and processing software required for the Stereo Particle Image Velocimetry (SPIV) system. Prior to the delivery of this equipment, scheduled for March 1998, the two aforementioned workstations will be sent to Dantech for installation of the data interface cards and software. Also this year, the ALE3D (Arbitrary Lagrangian Eulerian in 3D) code will be installed on an additional workstation to address coupled aerodynamic and structural modeling of problems in High Cycle Fatigue. Finally, the fourth workstation will be incorporated into the control system for active control studies.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	1
High-end Workstation	1
Low-end Server	0
High-end Server	0

Current Status

Substantial progress has been made during the past year for high fidelity, real time nuclear reactor simulation on the Quad Pentium Pro donated by Intel for the Technology for Education 2000 program. The nuclear reactor simulator code has been ported from Solaris to the Windows NT operating system using DIGITAL Visual Fortran. Results from a suite of standard benchmark problems using the NT-based simulator were then compared with benchmark results in order to verify the ported code.

Parallel implementation of the code was performed using an incomplete domain decomposition algorithm and directives from the Kuck and Associates (KAI) Application Simulator/Guide software package. In the Guide application parallel directives are embedded within Fortran code, and a pre-processor replaces these directives with appropriate calls to the system thread library. The Fortran compiler then generates an executable file that can be run on multiple processors such as the Pentium Pro Quad. The simulator program with the Guide directives was executed on a single processor, and results were again verified. However, execution on more than one processor yielded results that differed slightly from the benchmark results using multiple processors. An effort is currently underway to understand these differences.

Planned Activities

The first objective of the project for the next year will be to resolve differences in the serial and parallel solutions. A complete suite of reactor spatial kinetics benchmark problems will then be analyzed to verify the performance of the simulator. Once identical results are obtained across all versions of the simulator, reductions in the execution time will become the focal point of the project. By the end of the current semester, we will be able to assess how close we are to achieving real time performance on the Pentium Pro Quad and what further trade-offs in model fidelity may be necessary. We anticipate the development of more complex and higher fidelity models for advanced reactors that may require the next generation Pentium Pro.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0
High-end Workstation	0
Low-end Server	0
High-end Server	0

Current Status

Work on the project is currently on or slightly ahead of schedule. Four systems have been delivered to date for purposes of becoming familiar with the personal computer (PC) environment for use as a research and educational development platform. Thus, during the last several months the major activity has been porting research, educational, and custom development software. In addition, we are developing a training program for rapid introduction of new graduate students to the use of our newly ported development tools on the PC. The initial systems have been used to found an Intel Computer Integrated Process Operations Laboratory (Intel CIPAC Laboratory) which will be the focal point of our industrial consortium's computing activities. The industry members of the consortium include Air Products, The Coca Cola Company, Dow Chemical, Dupont, Gensym, Lilly, Mitsubishi Chemical, Monsanto, Westvaco, and Weyerhaueser.

Planned Activities

Given the experience with the initial systems, our immediate plans are to expand the Intel CIPAC Laboratory for use by a larger proportion of the graduate students involved in consortium activities. By summer 1998, major development activities for research projects sponsored by the consortium will be conducted using Intel provided personal computers. By Fall 1998, we plan to replace our current workstations in the process systems educational laboratory with personal computers that will run the software that we are porting as well as other educational software packages used within our curriculum (e.g., Matlab, etc.). The process systems educational laboratory is used by most core undergraduate courses within our School.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	4	0	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

Current Status

(not available)

Planned Activities

(not available)

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

The objective of this project is to provide an environment whereby students and faculty can use the latest engineering design and analysis software coupled with CAD software and rapid prototyping equipment so that the design can be realized in hardware and then tested. A laboratory with Graphic Servers and Desktop Workstations is being set up in Grissom Hall. Physical Plant is currently installing the wiring. New furniture has been ordered with delivery and completion of the room scheduled for the end of February. A 4-axis CNC milling machine was purchased along with the necessary software and installed at the Aerospace Sciences Laboratory. Students are using the workstations that have been delivered with modern design/analysis software, such as NASTRAN, ANSYS, ABACUS, FLUENT, CDD2000, PMARC, IDEAS, etc., to design and analyze components of an aerospace vehicles.

Planned Activities

Current efforts are directed at interfacing the design laboratory in Grissom Hall with the machine tools at the Aerospace Sciences Laboratory. Purchase of additional rapid prototyping hardware is planned. All workstations will be moved to Grissom 380 when physical plant is finished.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	0	0
High-end Workstation	5	5	5	5
Low-end Server	0	0	0	0
High-end Server	0	2	0	2

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	3	0
High-end Workstation	0	0	1	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

Current Status

(not available)

Planned Activities

(not available)

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

(not available)

Planned Activities

(not available)

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation
High-end Workstation
Low-end Server
High-end Server

(no justification provided)

Current Status

All of the systems that have been delivered (15 "low-end" workstations, one dual-processor high-end workstation, two dual-processor servers, and one quad-processor server) are networked, on-line and actively being used. New video capture and production hardware/software is in place, and we are beginning to create new on-line course modules.   A prototype of the on-line lab manual has also been created, and is currently being tested. A shortage of disk drives for the large servers, however, is currently limiting the scope of what can be immediately accomplished.

Planned Activities

A major curriculum development effort is underway in the digital systems course sequence that is the focus of this project. Efforts specifically associated with this project include creating multimedia-delivered course content along with a new set of highly integrated laboratory exercises and design projects that are centered on a teamwork approach. Also, a significant amount of data relative to performance of the instructional multimedia testbed will be available during the first year of the project. Continuous curriculum development and enhancements, along with research into digital video compression/caching, will provide results of strategic interest to Intel during years two and three of this project.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	13	0	0
High-end Workstation	0	2	0	0
Low-end Server	0	0	0	0
High-end Server	0	0	0	0

The workstations requested will be used for the "middle" course in the digital systems curriculum being revamped (EE 3xx Microprocessor Systems and Interfacing).

Current Status

One quad processor system was requested and delivered. This system has been used to pilot Windows Application Serving to UNIX and Macintosh systems on campus. It has also been used to deploy heavyweight Windows NT applications to off campus student owned computers using Citrix WinFrame and Insignia NTrigue. It has also been used for testing and development of X.500 and LDAP directory services for the IMAP mail project. Users have also been testing "number crunch" research applications on this server in a multiuser mode.

More information is available in a summary presentation.

Planned Activities

Four desktop systems were requested and delivered. The monitors for these systems still have not arrived. With borrowed monitors, 3 of these systems are serving as desktop systems for PACE staff and are primarily used for testing of experimental configurations. The fourth system will serve as a test WinFrame Picasso/Windows Hydra system and will also provide NT server services to the PACE staff NT workstations. A key goal of the next quarter is to evaluate the scalability of the quad processor application server.

Estimated Number of Systems to be Requested in 1998

	98Q1	98Q2	98Q3	98Q4
Low-end Workstation	0	0	2	1
High-end Workstation	0	0	0	0
Low-end Server	0	0	1	1
High-end Server	0	0	1	0

In year two, we will be anxious for arrival of the equipment in order to be able to keep accomplishment timely for PACE goals. Thus, we request all year 2 equipment in the first two quarters of year 2; that is, Q3 and Q4, 1998.