Mesh traversal and sorting for efficient memory usage in scientific codes

Applications that operate on meshes are very popular in High Performance Computing (HPC) environments. In the past, many techniques have been developed in order to optimize the memory accesses for these datasets. Different loop transformations and domain decompositions are com- monly used for structured meshes. However, unstructured grids are more challenging. The memory accesses, based on the mesh connectivity, do not map well to the usual lin- ear memory model. This work presents a method to improve the memory performance which is suitable for HPC codes that operate on meshes. We develop a method to adjust the sequence in which the data are used inside the algorithm, by means of traversing and sorting the mesh. This sorted mesh can be transferred sequentially to the lower memory levels and allows for minimum data transfer requirements. The method also reduces the lower memory requirements dra- matically: up to 63% of the L1 cache misses are removed in a traditional cache system. We have obtained speedups of up to 2.58 on memory operations as measured in a general- purpose CPU. An improvement is also observed with se- quential access memories, where we have observed reduc- tions of up to 99% in the required low-level memory size.

Review and comparison of effective delayed neutron fraction calculation methods with Monte Carlo codes

The calculation of the effective delayed neutron fraction, beff , with Monte Carlo codes is a complex task due to the requirement of properly considering the adjoint weighting of delayed neutrons. Nevertheless, several techniques have been proposed to circumvent this difficulty and obtain accurate Monte Carlo results for beff without the need of explicitly determining the adjoint flux. In this paper, we make a review of some of these techniques; namely we have analyzed two variants of what we call the k-eigenvalue technique and other techniques based on different interpretations of the physical meaning of the adjoint weighting. To test the validity of all these techniques we have implemented them with the MCNPX code and we have benchmarked them against a range of critical and subcritical systems for which either experimental or deterministic values of beff are available. Furthermore, several nuclear data libraries have been used in order to assess the impact of the uncertainty in nuclear data in the calculated value of beff .

State of the Art and Future Trends in Grid Codes Applicable to Isolated Electrical Systems

Isolated electrical systems lack electrical interconnection to other networks and are usually placed in geographically isolated areas—mainly islands or locations in developing countries. Until recently, only diesel generators were able to assure a safe and reliable supply in exchange for very high costs for fuel transportation and system operation. Transmission system operators (TSOs) are increasingly seeking to replace traditional energy models based on large groups of conventional generation units with mixed solutions where diesel groups are held as backup generation and important advantages are provided by renewable energy sources. The grid codes determine the technical requirements to be fulfilled by the generators connected in any electrical network, but regulations applied to isolated grids are more demanding. In technical literature it is rather easy to find and compare grid codes for interconnected electrical systems. However, the existing literature is incomplete and sparse regarding isolated grids. This paper aims to review the current state of isolated systems and grid codes applicable to them, specifying points of comparison and defining the guidelines to be followed by the upcoming regulations.

Shake-table tests of a reinforced concrete frame designed following modern codes: seismic performance and damage evaluation

This paper presents shake-table tests conducted on a two-fifths-scale reinforced concrete frame representing a conventional construction design under current building code provisions in the Mediterranean area. The structure was subjected to a sequence of dynamic tests including free vibrations and four seismic simulations in which a historical ground motion record was scaled to levels of increasing intensity until collapse. Each seismic simulation was associated with a different level of seismic hazard, representing very frequent, frequent, rare and very rare earthquakes. The structure remained basically undamaged and within the inter-story drift limits of the "immediate occupancy" performance level for the very frequent and frequent earthquakes. For the rare earthquake, the specimen sustained significant damage with chord rotations of up to 28% of its ultimate capacity and approached the upper bound limit of inter-story drift associated with "life safety". The specimen collapsed at the beginning of the "very rare" seismic simulation. Besides summarizing the experimental program, this paper evaluates the damage quantitatively at the global and local levels in terms of chord rotation and other damage indexes, together with the energy dissipation demands for each level of seismic hazard. Further, the ratios of column-to-beam moment capacity recommended by Eurocode 8 and ACI-318 to guarantee the formation of a strong column-weak beam mechanism are examined.

State of the Art and Future Trends in Grid Codes Applicable to Isolated Electrical Systems

Electrical power systems are changing their traditional structure, which was based on a little number of large generating power plants placed at great distances from loads by new models that tend to split the big production nodes in many smaller ones. The set of small groups which are located close to consumers and provide safe and quality energy is called distributed generation (DG). The proximity of the sources to the loads reduces losses associated with transportation and increases overall system efficiency. DG also favors the inclusion of renewable energy sources in isolated electrical systems or remote microgrids, because they can be installed where the natural resource is located. In both cases, as weak grids unable to get help from other nearby networks, it is essential to ensure appropriate behavior of DG sources to guarantee power system safety and stability. The grid codes sets out the technical requirements to be fulfilled for the sources connected in these electrical networks. In technical literature it is rather easy to find and compare grid codes for interconnected electrical systems. However, the existing literature is incomplete and sparse regarding isolated electrical systems and this happens due to the difficulties inherent in the pursuit of codes. Some countries have developed their own legislation only for their island territory (as Spain or France), others apply the same set of rules as in mainland, another group of island countries have elaborated a complete grid code for all generating sources and some others lack specific regulation. This paper aims to make a complete review of the state of the art in grid codes applicable to isolated systems, setting the comparison between them and defining the guidelines predictably followed by the upcoming regulations in these particular systems.

Efficient information reconciliation for Continuous-Variable QKD using Non-Binary Low-Density Parity-Check Codes

We present here an information reconciliation method and demonstrate for the first time that it can achieve efficiencies close to 0.98. This method is based on the belief propagation decoding of non-binary LDPC codes over finite (Galois) fields. In particular, for convenience and faster decoding we only consider power-of-two Galois fields.

Interleaved concatenated codes: New perspectives on approaching the Shannon limit

The last few years have witnessed a significant decrease in the gap between the Shannon channel capacity limit and what is practically achievable. Progress has resulted from novel extensions of previously known coding techniques involving interleaved concatenated codes. A considerable body of simulation results is now available, supported by an important but limited theoretical basis. This paper presents a computational technique which further ties simulation results to the known theory and reveals a considerable reduction in the complexity required to approach the Shannon limit.

Synthetic genes for glycoprotein design and the elucidation of hydroxyproline-O-glycosylation codes

Design of hydroxyproline (Hyp)-rich glycoproteins (HRGPs) offers an approach for the structural and functional analysis of these wall components, which are broadly implicated in plant growth and development. HRGPs consist of multiple small repetitive “glycomodules” extensively O-glycosylated through the Hyp residues. The patterns of Hyp-O-glycosylation are putatively coded by the primary sequence as described by the Hyp contiguity hypothesis, which predicts contiguous Hyp residues to be attachment sites of small arabinooligosaccharides (1–5 Ara residues/Hyp); while clustered, noncontiguous Hyp residues are sites of arabinogalactan polysaccharide attachment. As a test, we designed two simple HRGPs as fusion proteins with green fluorescent protein. The first was a repetitive Ser-Hyp motif that encoded only clustered noncontiguous Hyp residues, predicted polysaccharide addition sites. The resulting glycoprotein had arabinogalactan polysaccharide O-linked to all Hyp residues. The second construct, based on the consensus sequence of a gum arabic HRGP, contained both arabinogalactan and arabinooligosaccharide addition sites and, as predicted, gave a product that contained both saccharide types. These results identify an O-glycosylation code of plants.

The affected gene underlying the class K glycosylphosphatidylinositol (GPI) surface protein defect codes for the GPI transamidase

The final step in glycosylphosphatidylinositol (GPI) anchoring of cell surface proteins consists of a transamidation reaction in which preassembled GPI donors are substituted for C-terminal signal sequences in nascent polypeptides. In previous studies we described a human K562 cell mutant, termed class K, that accumulates fully assembled GPI units but is unable to transfer them to N-terminally processed proproteins. In further work we showed that, unlike wild-type microsomes, microsomes from these cells are unable to support C-terminal interaction of proproteins with the small nucleophiles hydrazine or hydroxylamine, and that the cells thus are defective in transamidation. In this study, using a modified recombinant vaccinia transient transfection system in conjunction with a composite cDNA prepared by 5′ extension of an existing GenBank sequence, we found that the genetic element affected in these cells corresponds to the human homolog of yGPI8, a gene affected in a yeast mutant strain exhibiting similar accumulation of GPI donors without transfer. hGPI8 gives rise to mRNAs of 1.6 and 1.9 kb, both encoding a protein of 395 amino acids that varies in cells with their ability to couple GPIs to proteins. The gene spans ≈25 kb of DNA on chromosome 1. Reconstitution of class K cells with hGPI8 abolishes their accumulation of GPI precursors and restores C-terminal processing of GPI-anchored proteins. Also, hGPI8 restores the ability of microsomes from the mutant cells to yield an active carbonyl in the presence of a proprotein which is considered to be an intermediate in catalysis by a transamidase.

A single gene of chloroplast origin codes for mitochondrial and chloroplastic methionyl–tRNA synthetase in Arabidopsis thaliana

One-fifth of the tRNAs used in plant mitochondrial translation is coded for by chloroplast-derived tRNA genes. To understand how aminoacyl–tRNA synthetases have adapted to the presence of these tRNAs in mitochondria, we have cloned an Arabidopsis thaliana cDNA coding for a methionyl–tRNA synthetase. This enzyme was chosen because chloroplast-like elongator tRNAMet genes have been described in several plant species, including A. thaliana. We demonstrate here that the isolated cDNA codes for both the chloroplastic and the mitochondrial methionyl–tRNA synthetase (MetRS). The protein is transported into isolated chloroplasts and mitochondria and is processed to its mature form in both organelles. Transient expression assays using the green fluorescent protein demonstrated that the N-terminal region of the MetRS is sufficient to address the protein to both chloroplasts and mitochondria. Moreover, characterization of MetRS activities from mitochondria and chloroplasts of pea showed that only one MetRS activity exists in each organelle and that both are indistinguishable by their behavior on ion exchange and hydrophobic chromatographies. The high degree of sequence similarity between A. thaliana and Synechocystis MetRS strongly suggests that the A. thaliana MetRS gene described here is of chloroplast origin.

Class I release factors in ciliates with variant genetic codes

In eukaryotes with the universal genetic code a single class I release factor (eRF1) most probably recognizes all stop codons (UAA, UAG and UGA) and is essential for termination of nascent peptide synthesis. It is well established that stop codons have been reassigned to amino acid codons at least three times among ciliates. The codon specificities of ciliate eRF1s must have been modified to accommodate the variant codes. In this study we have amplified, cloned and sequenced eRF1 genes of two hypotrichous ciliates, Oxytricha trifallax (UAA and UAG for Gln) and Euplotes aediculatus (UGA for Cys). We also sequenced/identified three protist and two archaeal class I RF genes to enlarge the database of eRF1/aRF1s with the universal code. Extensive comparisons between universal code eRF1s and those of Oxytricha, Euplotes and Tetrahymena, which represent three lineages that acquired variant codes independently, provide important clues to identify stop codon-binding regions in eRF1. Domain 1 in the five ciliate eRF1s, particulary the TASNIKS heptapeptide and its adjacent region, differs significantly from domain 1 in universal code eRF1s. This observation suggests that domain 1 contains the codon recognition site, but that the mechanism of eRF1 codon recognition may be more complex than proposed by Nakamura et al. or Knight and Landweber.

Ensemble codes involving hippocampal neurons are at risk during delayed performance tests

Multielectrode recording techniques were used to record ensemble activity from 10 to 16 simultaneously active CA1 and CA3 neurons in the rat hippocampus during performance of a spatial delayed-nonmatch-to-sample task. Extracted sources of variance were used to assess the nature of two different types of errors that accounted for 30% of total trials. The two types of errors included ensemble “miscodes” of sample phase information and errors associated with delay-dependent corruption or disappearance of sample information at the time of the nonmatch response. Statistical assessment of trial sequences and associated “strength” of hippocampal ensemble codes revealed that miscoded error trials always followed delay-dependent error trials in which encoding was “weak,” indicating that the two types of errors were “linked.” It was determined that the occurrence of weakly encoded, delay-dependent error trials initiated an ensemble encoding “strategy” that increased the chances of being correct on the next trial and avoided the occurrence of further delay-dependent errors. Unexpectedly, the strategy involved “strongly” encoding response position information from the prior (delay-dependent) error trial and carrying it forward to the sample phase of the next trial. This produced a miscode type error on trials in which the “carried over” information obliterated encoding of the sample phase response on the next trial. Application of this strategy, irrespective of outcome, was sufficient to reorient the animal to the proper between trial sequence of response contingencies (nonmatch-to-sample) and boost performance to 73% correct on subsequent trials. The capacity for ensemble analyses of strength of information encoding combined with statistical assessment of trial sequences therefore provided unique insight into the “dynamic” nature of the role hippocampus plays in delay type memory tasks.

Neural codes: Firing rates and beyond

Computational neuroscience has contributed significantly to our understanding of higher brain function by combining experimental neurobiology, psychophysics, modeling, and mathematical analysis. This article reviews recent advances in a key area: neural coding and information processing. It is shown that synapses are capable of supporting computations based on highly structured temporal codes. Such codes could provide a substrate for unambiguous representations of complex stimuli and be used to solve difficult cognitive tasks, such as the binding problem. Unsupervised learning rules could generate the circuitry required for precise temporal codes. Together, these results indicate that neural systems perform a rich repertoire of computations based on action potential timing.

A unique human gene that spans over 230 kb in the human chromosome 8p11-12 and codes multiple family proteins sharing RNA-binding motifs.

A unique gene, RBP-MS, spanning over 230 kb in the human chromosome 8p11-12 near the Werner syndrome gene locus is described. The single-copy RBP-MS gene is alternatively spliced, resulting in a family of at least 12 transcripts (average length of 1.5 kb). Nine different types of cDNAs that encode an RNa-binding motif at the N terminus and helix-rich sequences at the C terminus have been identified thus far. Among the 16 exons identified, four 5'-proximal exons contained sequences homologous to the RNA-binding domain of Drosophila couch potato gene. Northern blot analysis showed that the RBP-MS gene was expressed strongly in the heart, prostate, intestine, and ovary, and poorly in the skeletal muscle, spleen, thymus, brain, and peripheral leukocytes. The possible role of this gene in RNA metabolism is discussed.

Cloning and expression in Escherichia coli of the OGG1 gene of Saccharomyces cerevisiae, which codes for a DNA glycosylase that excises 7,8-dihydro-8-oxoguanine and 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine.

A spontaneous mutator strain of Escherichia coli (fpg mutY) was used to clone the OGG1 gene of Saccharomyces cerevisiae, which encodes a DNA glycosylase activity that excises 7,8-dihydro-8-oxoguanine (8-OxoG). E. coli (fpg mutY) was transformed by a yeast DNA library, and clones that showed a reduced spontaneous mutagenesis were selected. The antimutator activity was associated with pYSB10, an 11-kbp recombinant plasmid. Cell-free extracts of E. coli (fpg mutY) harboring pYSB10 possess an enzymatic activity that cleaves a 34-mer oligonucleotide containing a single 8-oxoG opposite a cytosine (8-OxoG/C). The yeast DNA fragment of 1.7 kbp that suppresses spontaneous mutagenesis and overproduces the 8-OxoG/C cleavage activity was sequenced and mapped to chromosome XIII. DNA sequencing identified an open reading frame, designated OGG1, which encodes a protein of 376 amino acids with a molecular mass of 43 kDa. The OGG1 gene was inserted in plasmid pUC19, yielding pYSB110. E. coli (fpg) harboring pYSB110 was used to purify the Ogg1 protein of S. cerevisiae to apparent homogeneity. The Ogg1 protein possesses a DNA glycosylase activity that releases 8-OxoG and 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine. The Ogg1 protein preferentially incises DNA that contains 8-OxoG opposite cytosine (8-OxoG/C) or thymine (8-OxoG/T). In contrast, Ogg1 protein does not incise the duplex where an adenine is placed opposite 8-OxoG (8-OxoG/A). The mechanism of strand cleavage by Ogg1 protein is probably due to the excision of 8-OxoG followed by a beta-elimination at the resulting apurinic/apyrimidinic site.