Sadegh Davari, Ph.D., Associate Professor; Theodore Leibfried, Ph.D., Associate Professor; Kwok-Bun Yue, Ph.D., Associate Professor, Natural and Applied Sciences, UHCL
Investigators studied several research issues concerning scheduling real-time tasks, particularly programs on multi-processors and distributed systems.
Whereas others have focused on satisfying one or two selected requirements in task scheduling for multi-processor systems, our research team identified the major requirements common to most practical real-time systems. These requirements include the existence of aperiodic tasks, task synchronization, fault tolerance and the existence of task suspensions.
Several existing solutions on generalizing task scheduling algorithms to handle aperiodic tasks and task synchronization bore investigation and were compared to the results of the algorithms we developed.1
An initial framework has been identified for classifying solutions for task scheduling in real-time multi-processor systems. This framework will be used to classify existing and future solutions.
The results of this investigation have resulted in the submission of a major proposal to the National Science Foundation.2 This external proposal seeks the development of a comprehensive solution for task scheduling in practical real-time multi-processors that can handle the prescribed requirements previouslymentioned. A side-product will be a systematic classification of task scheduling solutions in multi-processors, a survey, and analysis much needed by the real-time research community.
Three additional important results of this project can be identified:
References
1Davari, S. and S. K. Dhall. "On-Line Algorithms for Allocating
Periodic-Time-Critical Tasks on Multiprocessor Systems." Int'l J. Informatica 19
(1995): 83-96.
2Davari, S., K. Yue, and T. Leibfried. "Guaranteeing Hard Deadline of Tasks in Future
Multi-Processor Systems." National Science Foundation, Nov. 1996. (Submitted.)
3Wang, Z., S. Davari, G. Collins, and W. Zhao. "Asynchronous Mode of FDDI for
Real-Time Multimedia Communication." Proc. 4th Int'l Workshop on Parallel and
Distributed Real-Time Sys., Honolulu, HA, April 15-16, 1996.
4Yue, K. and S. Davari. "Task Synchronization in Distributed Systems." High
Technology Laboratory Jan. 1996, $7,900.
5Yue, K., S. Davari, and T. Leibfried. "Priority Ceiling Protocol in Ada-95."
TriAda Conf., 1996. (Submitted for publication.)
Back to Table of Contents
William E. Fitzgibbon, III, Ph.D., Professor, and Roland Glowinski, Ph.D., Professor, Mathematics, UH
The University of Houston, with the collaboration of INRIA/Sophia Antipolis, hosted a two day conference November 6-9 on the experimental and computational understanding of high speed flows (both supersonic and hypersonic flows) followed by a workshop. INRIA is the Institute Nationale de Recherche en Informatique et en Automatique, the center for computer research in France.
Events held on the University of Houston campus consisted of sixteen invited plenary lectures delivered by recognized experts of the state-of-the-art regarding the experimental and computational understanding of high speed flows (both supersonic and hypersonic flows). Plenary lectures were accompanied by a poster session which comprised thirty poster presentations.
The lectures and poster presentations featured and developed the following scientific themes and technological highlights:
Scientific Themes
Technological Highlights
The extremity of physical conditions severely restricts the availability of experimental data for high speed flows. Moreover, the modeling difficulties are profound. Therefore, detailed understanding and accurate prediction of high speed flow phenomena require a combination of experimental and computational methodologies. The workshop aimed at comparing experimental data with computational simulations of eleven test cases. Test cases were selected by a panel of experts on both their relevance and the high quality of the supporting experimental results. At the conclusion of the workshop, syntheses utilizing Silicon Graphics Indigo 2 Workstations allowed the real-time visualization of computed solutions to test problems and evaluate them by comparison with experimental results.
Organizers were gratified by the uniform high quality of the oral and poster presentations and by the exceptional quality of the experimental results.
Among 85 participants, 40 were from the United States, 29 from Western Europe, 6 from Japan, and 6 from Russia.
Results of the workshop will be entered into the High Speed Flow Field Database located in the Mathematics De-partment. This on-line database will be maintained for later computational and experimental validation as well for code calibration. It will be well documented and easily accessible to the research and industrial communities. Documen-tation will include written reports, books, and an online browser. Eventually the database will be stored on a devoted disk attached to an Internet accessible mode. Backup tapes will be maintained by the University of Houston. Information stored on the European Hypersonic Database will also be stored on the HSFF Database.
The database will continue to accept new solutions. Criteria for acceptance will include accuracy, robustness, originality and, perhaps, prior unseen phenomena. Moreover, additional test problems will be formulated and solutions solicited. In this sense the database will function as a continuing reference bank and an electronic journal. Plans are under way for a follow up workshop.
Information on the conference and future events is available via the Internet at http://math.uh.edu/~hsff/.
Back to Table of Contents
Karen F. Frasier-Scott, Ph.D., Associate Professor, and Sherry A. Bergh, M.S. School of Natural and Applied Science, UHCL
Progress in the last decade has led us toward a greater understanding of the influence of gravitational forces on mammalian cell growth, differentiation, and affects on nuclear transport. The regulation of nuclear transport affects DNA replication gene expression and protein synthesis. Translocation of mRNA through the nuclear envelope pore complex does not occur passively but is known to require specific RNA recognition signals, nuclear envelope binding proteins, and enzymes. Growth factors, hormones, protein kinases, and phosphoprotein phosphatases may be involved in nuclear mRNA transport. Mechanical forces may influence these recognition signals and change the nuclear pore size and transport function through messages that are transmitted by the cytoskeleton. Structural/functional aspects of the cytoskeleton may also influence nuclear transport.
The endothelium plays a pivotal role in modulating a number of physiologic processes including hemostasis, thrombosis, inflammation, and immune responses. Endothelial cell monolayers are interconnected by tight junctions, adherence junctions, and gap junctions. The integrity of the monolayer is maintained by these cell-cell junctional complexes, together with a wide spectrum of adhesion molecules coupled to the endothelial cytoskeleton. The cytoskeleton plays a crucial role in maintaining normal cell morphology and mediating intra- and extracellular responses. To elucidate the role of the cytoskeleton, externally applied mechanical forces such as fluid shear stress and flexing mechanisms have been utilized to perturb a cell's morphology and cytoskeletal elements.
Figure 1. In vitro
studies demonstrated that when the cells were exposed to long- term hypergravity at 30xg,
both the total ACE-mRNA and ACE-hnRNA expression increased slightly in comparison to
controls in pre-confluent cells.
Angiotensin converting enzyme (ACE) is a dipeptidyl carboxypeptidase known to convert Angiotensin I to the vasoactive Angiotensin 11 and to hydrolyze and inactivate the vasodilator bradykinin. The endothelial form of ACE is a component of the renin-angiotensin system (RAS) and is thought to play a significant role in blood pressure regulation and in fluid and electrolytic balance. ACE enzyme activity is modulated in vitro by numerous chemical agents as well as by cell culture conditions. The mechanism of action of these modulators on ACE is poorly understood. Yet even less is known about the in vivo regulation of ACE-mRNA and the expression of ACE-protein in endothelial cells.
We have used the RT/PCR in combination with digoxigenin labeled probes for in situ hybridization to provide a sensitive detection method for Angiotensin converting enzyme mRNA (ACE-mRNA) and heteronuclear RNA (ACE-hnRNA) expression in BAECs in culture. Since mRNAs are transient intermediates, their relative concentration at any given time reflects their rates of formation (transcription) and/or disappearance (degradation). Therefore, relative increases in mRNA reflect either increases in transcription and/or decreases in the rate of degradation of the mRNA. Thus, RT/PCR detection of mRNA provides information on transcriptional activity indirectly by measuring the relative steady state amount of a transient intermediate that is produced by transcription while in situ studies provide direct information about the transcriptional activity of ACE-mRNA and ACE-hnRNA.
Figure 2. When the cells
were exposed to long term hypergravityat 35xg, increases were still seen in the total
ACE-hnRNA population, but the total ACE-mRNA concentraton decreased markedly with time.
We have used the Bovine Aortic Endothelial Cells (BAECs) for both an in situ and in vitro model system for nuclear transport to localize Angiotensin Converting enzyme-RNA (ACE-RNA) within cycling and confluent cells in a hypergravity environment. BAECs were cultured in variable hypergravity using the Savant tissue culture centrifuge (Savant, Farmingdale, NY). Control experiments were maintained at 1xg under standard conditions. In situ and in vitro hybridizations were performed using both the 480 bp digoxigenin labeled ACE exon probe and the 190 bp intron/exon ACE probe during parallel experiments. In situ studies demonstrated that pre-confluent endothelial cells exposed to short term hypergravity (24 hrs) initially increased the expression of both ACE-hnRNA and ACE-hnRNA within the nucleus when compared to controls. By increasing the time that the cells are exposed to hypergravity at 10xg and 35xg, we could detect a marked decrease in nuclear expression of both ACE-hnRNA and ACE-hnRNA in comparison with controls. In vitro studies demonstrated that when the cells were exposed to long term hypergravity at 30xg, both the total ACE-mRNA and ACE-hnRNA expression (Fig. 1) increased slightly in comparison to controls in pre-confluent cells. When the cells were exposed to long term hypergravity at 35xg (Fig. 2), increases were still seen in the total ACE-hnRNA population, but the total ACE-mRNA concentration decreased markedly with time. Parallel experiments demonstrated that when cells were exposed to long term hypergravity at 30xg and 35xg (Fig. 3), their rate of growth increased dramatically at all time points tested in comparison to controls.
We have developed two very important tools, RT/PCR and in situ hybridization, to document ACE gene expression in endothelial cells. With Probes designed to detect both ACE-mRNA and ACE-hnRNA, we are able to follow and compare the nuclear transport and maturation process of ACE expression during exponential growth of endothelial cells exposed to hypergravity.
Figure 3. Pre-confluent
variable hypergravity cell counts. When cells were exposed to long term hypergravity at
30xg and 35xg, their rate of growth increased dramatically at all time points tested in
comparison to controls.
Contents
ISSO -- Institute for Space Systems Operations
1994-1995 Annual Report
|
|