1999 — 2003 |
Whalley, David [⬀] Yuan, Xin (co-PI) [⬀] Van Engelen, Robert |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Automatic Validation of Improving Transformations and Related Applications @ Florida State University
Proposal Number: CCR-9904943
Title: Automatic Validation of Code Improving Transformations and Related Applications
Investigators: Robert van Engelen, David Whalley, and Xin Yuan
Abstract:
Ensuring that software systems execute correctly is vital. One portion of this process is to ensure that the compiler produces machine code that accurately represents the algorithms specified at the source code level. This is a formidable task since an optimizing compiler not only translates the source code to machine code, it may apply hundreds or thousands of code improving transformations to even a relatively small program. This project will use two approaches that dramatically simplify the validation of code improving transformations. First, the equivalence of the program representation before and after each improving transformation will be checked, rather than attempting to equate the source and object programs directly. Each individual code improving transformation typically consists of only a few changes. Second, only the region of the program that has changed, as opposed to the entire program representation, will be checked. The regions of the program representation that are changed by an improving transformation are typically quite small. The equivalence will be shown by demonstrating that the effects of the region on the rest of the program will remain the same before and after the improving transformation. The benefits include simpler diagnosis of compiler errors and more reliable compilers.
|
0.915 |
2001 — 2005 |
Gallivan, Kyle (co-PI) [⬀] Van Engelen, Robert |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Improving Symbolic Analysis of Restructuring Compilers @ Florida State University
The proposed research investigates new methods for symbolic analysis to improve various restructuring compiler optimizations. A new algebra on functions is investigated to manipulate, simplify, and derive normal forms of scalar functions and (generalized) induction variables in multi-dimensional loops. The derivation of normal forms for intermediate program constructs enables reasoning about the semantics of a program under analysis. This is extremely useful to improve various compiler optimizations to effectively deal with symbolic expressions in real-world applications. More specifically, the proposed research aims to improve symbolic analysis methods such as generalized induction variable recognition, linear and non-linear data dependence analysis, value range analysis, global value propagation, and counting the number of solutions to systems of constraints. The effectiveness of parallelizing compilers depends heavily on the accuracy of these methods. The research will result in the ability of compilers to more effectively handle symbolic expressions and constraints. Current methods are not always effective, resulting in considerable performance losses caused by worst-case assumptions or when program analysis has to be performed at execution time.
|
0.915 |
2002 — 2006 |
Whalley, David [⬀] Yuan, Xin (co-PI) [⬀] Van Engelen, Robert |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Collaborative Research: Branch Elimination by Condition Merging @ Florida State University
Conditional branches are expensive. Branches require a significant percentage of the execution cycles since they occur frequently and can cause pipeline stalls. In addition, branches result in forks in the control flow, which can prevent other code-improving transformations from being applied. We plan to develop path profile-based techniques for replacing the execution of a set of two or more branches with a single branch on a conventional scalar processor. We propose to improve performance by merging the conditions of two or more branches into a single condition. Previous approaches have accomplished such merging of conditions that have either only involved a single variable or have required special hardware to merge multiple conditions together. Techniques will be developed to produce a merged condition involving multiple variables that can be used to bypass the code testing the original set of conditions on a conventional processor. Merging conditions may be very good fit for run-time optimization systems, which optimize frequently executed paths during the execution of a program.
|
0.915 |
2005 — 2009 |
Riccardi, Gregory (co-PI) [⬀] Erickson, Gregory (co-PI) [⬀] Van Engelen, Robert Mast, Austin Ronquist, Fredrik |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Morphbank: Web Image Database Technology For Comparative Morphology and Biodiversity Research @ Florida State University
Florida State University is granted an award to enhance the MorphBank Image database. Many biological disciplines rely for their data primarily on observations of appropriately prepared specimens. A few examples include comparative anatomy, histology, morphological phylogenetics, and taxonomic research. The need for access to suitable specimens constitutes a significant hurdle to rapid progress in these disciplines. However, in many cases data of comparable quality can be generated from photographs of the specimens instead of from the specimens themselves. The development of fast Internet connections and the availability of low-cost high-quality digital imaging techniques is currently revolutionizing research and teaching in these disciplines, and recent years have seen a rapid proliferation of 2D- and 3D-image databases of biological specimens on the web. MorphBank (www.morphbank.net), not to be confused with MorphoBank, is a searchable web repository of images for comparative morphology and biodiversity research, originally conceived by a consortium of Swedish, Spanish and American systematic entomologists. It currently contains about 6,000 images documenting morphological phylogenetics analyses, newly described taxa, and voucher specimens tied to GenBank sequences. MorphBank has been open to the user community for uploading of images since early 2003 and works well as a searchable image archive. The next steps in the development of specimen image databases focus on three projects, each exemplifying one important use case for such databases: (1) large phylogenetics projects; (2) remote annotation of natural history collections; and (3) comparative histology projects. The interdisciplinary research team, which includes biologists, computer scientists, a mathematician, and information retrieval specialists, will develop general technology that can easily be adapted to address the specific needs of each of these use cases, facilitating the incorporation of additional use cases in the future. In particular, the team will develop community annotation technology, making it possible for researchers to link their observations to specific images, without compromising the original image, and enabling their colleagues to search, process and comment on these annotations. The team will also develop the MorphBank user interface including a new image-handling client, develop networking technology to support database functionality, substantially revise and refine the table structure, and implement more sophisticated data filtering and backup technologies. Finally, the team aims to build an international network of institutions supporting MorphBank or MorphBank-compatible image databases and to spur the development of international standards for biological image databases. The MorphBank project provides ample opportunities for supporting educational and outreach activities through its open access to biological images and associated information.
|
0.915 |
2006 — 2008 |
Yuan, Xin [⬀] Van Engelen, Robert Gopalan, Kartik (co-PI) [⬀] |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Cri: Acquisition of An Infiniband Cluster With Smp Nodes @ Florida State University
Abstract
Program: NSF 04-588 CISE Computing Research Infrastructure Proposal: CNS 0551555 PI: Yuan, Xin Institution: Florida State University Title: CRI: Acquisition of an Infiniband Cluster with SMP Nodes
Florida State University investigators will acquire an Infiniband cluster computer with dual processor, 64-bit SMP nodes. The cluster, which can be configured to resemble various cluster architectures, will give researchers access to state-of-the-art processor, network, and cluster technologies. By providing contemporary platforms for software development, testing, and performance evaluation, the acquisition can substantially improve the output of the participating research groups, increase the visibility of their research, encourage industry participation in research, and enhance education. The projects that will benefit from the acquisition span a range of areas in distributed systems including operating systems, runtime systems and compilers, and applications. The projects are (1) optimizing MPI programs through compiled communication, (2) delayed finalization of MPI collective communication routines, (3) distributed cluster resource virtualization, (4) performance and scalability of web services, (5) an integrated approach for communication optimization in the cluster of workstations environment, and (6) exploring high performance distributed systems (educational project).
|
0.915 |
2007 — 2011 |
Gallivan, Kyle (co-PI) [⬀] Van Engelen, Robert |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Collaborative Research: Flow-Sensitive Program Analysis For Speculative Parallelization @ Florida State University
The rise of chip multiprocessors (CMP) featuring tens to hundreds of processing units on a single chip promises to significantly boost the performance of desktop systems, rivaling the performance of yesterday's supercomputers. Supercomputing applications typically exploit a high degree of parallelism present in the application's computational tasks, which allows multiple processing units to work on solving the problem simultaneously to obtain a solution fast. However, common software applications are not written for specialized supercomputer architectures and lack sufficient explicit exposure of parallelism to gain speedups from CMPs automatically. Therefore, a successful exploitation of CMPs requires a rethinking of design, coding, and debugging by application developers. Programming languages and program annotations that natively support parallel concepts will be increasingly more successful, as well as programming languages in which sequential code can be more easily converted into parallel code.
This research investigates the combination and enhancements of several successful approaches to expose more parallelism in program code automatically. Firstly, the investigators will merge flow- sensitive loop-variant variable detection and optimization with the chains of recurrences (CR) algebra together with the NLVI (nonlinear variable interval) test that is based on interval theory. This aims to reduce the number of false positives prohibiting parallelization of loops with array dependences. Secondly, techniques for speculative parallelization of loops will be enhanced with a new run-time dependence analysis algorithm based on the CR algebra, NLVI test, and the theory of axiomatic semantics. Thirdly, a set of program annotations will be introduced to support speculative parallelization. This benefits source-to-source compilers and programmers who can leverage these annotations to extract more parallelism from loops by exercising application-specific knowledge.
|
0.915 |
2013 — 2018 |
Whalley, David (co-PI) [⬀] Tyson, Gary Van Engelen, Robert Wang, An-I Zhang, Zhenghao (co-PI) [⬀] |
N/AActivity Code Description: No activity code was retrieved: click on the grant title for more information |
Broadening Participation in Computer Science @ Florida State University
This project is supporting two cohorts of financially disadvantaged students to address the shortage of labor, women, and underrepresented groups in computer science and related disciplines. Each scholarship is being awarded for four years to remove the financial barrier to degree completion. Sixteen S-STEM scholars in 2013 and 15 in 2014 are being selected based on (1) NSF citizenship/residence requirements, (2) financial need as determined by the FAFSA, (3) academic qualification, (4) personal qualities (e.g., leadership), and (5) contribution to the diversity of the computer science student body.
The project objectives include (1) identifying and recruiting students with interest and high potential in computer science, with priority given to women and underrepresented groups, (2) improving retention through cohort class enrollments, S-STEM tutor networks, and enriched interactions between S-STEM scholars and existing academic support units and student groups, (3) providing internship and research opportunities to S-STEM scholars for timely professional development, and (4) gathering feedback to refine the curriculum.
Intellectual Merit: This project is creating interactions with different campus units and examining financial, social, support structure, curriculum, and environmental factors that contribute to the lack of women and underrepresented groups in computer science. S-STEM cohorts are serving as focus groups to examine the computer science education at the university, including curriculum, counseling, mentoring, professional development, and extracurricular activities.
Broader Impacts: The scholarships are allowing financially disadvantaged and debt-averse students to focus more on their degrees. Recruitment activities focus on women and underrepresented students to reestablish their critical mass in the program. Project results and findings are being disseminated through a web portal and education publications.
|
0.915 |