Browsing by Author "Zhao, Yuan"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Array syntax compilation and performance tuning
    (2007) Zhao, Yuan; Kennedy, Ken
    Array syntax adds expressive power to a language by providing operations on and assignments to array sections. Thus it allows programmers to write clear and concise code. However, state-of-the-art vendor compilers fail to efficiently map array statements to underlying architectures for high performance. The inefficiency is caused by ineffectively solving the following three technical problems: (1) reducing the size of allocated temporary array; (2) extending solutions to the evolving architectures; (3) applying loop fusion to multiple array statements. Finding solutions to these problems is important because otherwise array syntax, though a high-level language feature, may not be widely used by application developers. To address the above problems, this research first develops a novel strategy that minimizes the allocated temporary arrays using loop alignment and loop skewing on scalar processors, thereby reducing memory traffic and improving cache utilization. It then extends the minimization strategy to exploit the increasing on-chip parallelism on evolving architectures that offer vector (e.g., SSE and AltiVec) and multi-core (e.g., CELL) capabilities. In addition, new techniques boost performance by improving data alignment and managing data movement, both of which are important on these new architectures. Last, this dissertation parameterizes loop fusion for performance tuning and explores the properties of the space of all possible loop fusion configurations, to expedite performance tuning of loop fusion for increasing data reuse across multiple array statements. These transformations and optimizations are implemented in a source-to-source research compiler with extensions to target short vector processors and CELL processor. Experiments show that array statements compiled with our strategy run as much as two times faster than those compiled directly by vendor compilers. Our exploration of loop fusion parameter space identifies good candidates for heuristic searching and space pruning, which are essential to make the performance tuning process practical. In summary, this dissertation demonstrates that advanced compilation techniques can significantly improve the performance of programs written in array syntax upon current state-of-the-art implementation across a variety of architectures, including the latest multi-core processors with vector capabilities.
  • Thumbnail Image
    Item
    Chiral plasmonics of self-assembled nanorod dimers
    (Nature Publishing Group, 2013) Ma, Wei; Kuang, Hua; Wang, Libing; Xu, Liguang; Chang, Wei-Shun; Zhang, Huanan; Sun, Maozhong; Zhu, Yinyue; Zhao, Yuan; Liu, Liqiang; Xu, Chuanlai; Link, Stephan; Kotov, Nicholas A.
    Chiral nanoscale photonic systems typically follow either tetrahedral or helical geometries that require four or more different constituent nanoparticles. Smaller number of particles and different chiral geometries taking advantage of the self-organization capabilities of nanomaterials will advance understanding of chiral plasmonic effects, facilitate development of their theory, and stimulate practical applications of chiroplasmonics. Here we show that gold nanorods self-assemble into side-by-side orientated pairs and ‘‘ladders’’ in which chiral properties originate from the small dihedral angle between them. Spontaneous twisting of one nanorod versus the other one breaks the centrosymmetric nature of the parallel assemblies. Two possible enantiomeric conformations with positive and negative dihedral angles were obtained with different assembly triggers. The chiral nature of the angled nanorod pairs was confirmed by 4p full space simulations and the first example of single-particle CD spectroscopy. Self-assembled nanorod pairs and ‘‘ladders’’ enable the development of chiral metamaterials, (bio)sensors, and new catalytic processes.
  • Thumbnail Image
    Item
    Genome-wide identification and functional analyses of calmodulin genes inﾠ Solanaceous ﾠspecies
    (BioMed Central, 2013) Zhao, Yuan; Liu, Wei; Xu, You-Ping; Cao, Jia-Yi; Braam, Janet; Cai, Xin-Zhong
    Background: Calmodulin (CaM) is a major calcium sensor in all eukaryotes. It binds calcium and modulates the activity of a wide range of downstream proteins in response to calcium signals. However, little is known about the CaM gene family in Solanaceous species, including the economically important species, tomato (Solanum lycopersicum), and the gene silencing model plant, Nicotiana benthamiana. Moreover, the potential function of CaM in plant disease resistance remains largely unclear. Results: We performed genome-wide identification of CaM gene families in Solanaceous species. Employing bioinformatics approaches, multiple full-length CaM genes were identified from tomato, N. benthamiana and potato (S. tuberosum) genomes, with tomato having 6 CaM genes, N. benthamiana having 7 CaM genes, and potato having 4 CaM genes. Sequence comparison analyses showed that three tomato genes, SlCaM3/4/5, two potato genes StCaM2/3, and two sets of N. benthamiana genes, NbCaM1/2/3/4 and NbCaM5/6, encode identical CaM proteins, yet the genes contain different intron/exon organization and are located on different chromosomes. Further sequence comparisons and gene structural and phylogenetic analyses reveal that Solanaceous species gained a new group of CaM genes during evolution. These new CaM genes are unusual in that they contain three introns in contrast to only a single intron typical of known CaM genes in plants. The tomato CaM (SlCaM) genes were found to be expressed in all organs. Prediction of cis-acting elements in 5' upstream sequences and expression analyses demonstrated that SlCaM genes have potential to be highly responsive to a variety of biotic and abiotic stimuli. Additionally, silencing of SlCaM2 and SlCaM6 altered expression of a set of signaling and defense-related genes and resulted in significantly lower resistance to Tobacco rattle virus and the oomycete pathogen, Pythium aphanidermatum. Conclusions: The CaM gene families in the Solanaceous species tomato, N. benthamiana and potato were identified through a genome-wide analysis. All three plant species harbor a small set of genes that encode identical CaM proteins, which may manifest a strategy of plants to retain redundancy or enhanced quantitative gene function. In addition, Solanaceous species have evolved one new group of CaM genes during evolution. CaM genes play important roles in plant disease resistance to a variety of pathogens.
  • Thumbnail Image
    Item
    Scalarizing Fortran 90 Array Syntax
    (2001-03-30) Kennedy, Ken; Zhao, Yuan
    Array syntax is an important feature introduced in Fortran 90. It adds more expressive power to the language by allowing operations and assignments on the array sections. Programmers will benefit from this new feature directly by writing simple and concise programs. The remaining work is left to compilers that finally compile these statements with array syntax onto targeted machines. This compiling process is usually called scalarization. One problem facing Fortran 90 compilers is the temporary arrays generated during the scalarization in order to maintain the semantics of the array syntax. For many years, techniques have been developed to avoid the use of temporary arrays in certain cases. In this paper, we will present two new approaches to improve the scalarization. Our approaches show that in certain cases, the temporary array can be eliminated or reduced, when exact dependence information can be obtained. By doing so, our approaches help improve the memory hierarchy performance on a single scalar machine, in terms of the memory bandwidth.