A benchmark-driven modelling approach for evaluating deployment choices on a multi-core architecture

The complexity of current and emerging architectures provides users with options about how best to use the available resources, but makes predicting performance challenging. In this work a benchmark-driven model is developed for a simple shallow water code on a Cray XE6 system, to explore how deployment choices such as domain decomposition and core affinity affect performance. The resource sharing present in modern multi-core architectures adds various levels of heterogeneity to the system. Shared resources often includes cache, memory, network controllers and in some cases floating point units (as in the AMD Bulldozer), which mean that the access time depends on the mapping of application tasks, and the core's location within the system. Heterogeneity further increases with the use of hardware-accelerators such as GPUs and the Intel Xeon Phi, where many specialist cores are attached to general-purpose cores. This trend for shared resources and non-uniform cores is expected to continue into the exascale era. The complexity of these systems means that various runtime scenarios are possible, and it has been found that under-populating nodes, altering the domain decomposition and non-standard task to core mappings can dramatically alter performance. To find this out, however, is often a process of trial and error. To better inform this process, a performance model was developed for a simple regular grid-based kernel code, shallow. The code comprises two distinct types of work, loop-based array updates and nearest-neighbour halo-exchanges. Separate performance models were developed for each part, both based on a similar methodology. Application specific benchmarks were run to measure performance for different problem sizes under different execution scenarios. These results were then fed into a performance model that derives resource usage for a given deployment scenario, with interpolation between results as necessary.

Joint multiple dictionary learning for tensor sparse coding

Traditional dictionary learning algorithms are used for finding a sparse representation on high dimensional data by transforming samples into a one-dimensional (1D) vector. This 1D model loses the inherent spatial structure property of data. An alternative solution is to employ Tensor Decomposition for dictionary learning on their original structural form —a tensor— by learning multiple dictionaries along each mode and the corresponding sparse representation in respect to the Kronecker product of these dictionaries. To learn tensor dictionaries along each mode, all the existing methods update each dictionary iteratively in an alternating manner. Because atoms from each mode dictionary jointly make contributions to the sparsity of tensor, existing works ignore atoms correlations between different mode dictionaries by treating each mode dictionary independently. In this paper, we propose a joint multiple dictionary learning method for tensor sparse coding, which explores atom correlations for sparse representation and updates multiple atoms from each mode dictionary simultaneously. In this algorithm, the Frequent-Pattern Tree (FP-tree) mining algorithm is employed to exploit frequent atom patterns in the sparse representation. Inspired by the idea of K-SVD, we develop a new dictionary update method that jointly updates elements in each pattern. Experimental results demonstrate our method outperforms other tensor based dictionary learning algorithms.

IntFOLD: an integrated server for modelling protein structures and functions from amino acid sequences

IntFOLD is an independent web server that integrates our leading methods for structure and function prediction. The server provides a simple unified interface that aims to make complex protein modelling data more accessible to life scientists. The server web interface is designed to be intuitive and integrates a complex set of quantitative data, so that 3D modelling results can be viewed on a single page and interpreted by non-expert modellers at a glance. The only required input to the server is an amino acid sequence for the target protein. Here we describe major performance and user interface updates to the server, which comprises an integrated pipeline of methods for: tertiary structure prediction, global and local 3D model quality assessment, disorder prediction, structural domain prediction, function prediction and modelling of protein-ligand interactions. The server has been independently validated during numerous CASP (Critical Assessment of Techniques for Protein Structure Prediction) experiments, as well as being continuously evaluated by the CAMEO (Continuous Automated Model Evaluation) project. The IntFOLD server is available at: http://www.reading.ac.uk/bioinf/IntFOLD/

Conditioning of the weak-constraint variational data assimilation problem for numerical weather prediction

4-Dimensional Variational Data Assimilation (4DVAR) assimilates observations through the minimisation of a least-squares objective function, which is constrained by the model flow. We refer to 4DVAR as strong-constraint 4DVAR (sc4DVAR) in this thesis as it assumes the model is perfect. Relaxing this assumption gives rise to weak-constraint 4DVAR (wc4DVAR), leading to a different minimisation problem with more degrees of freedom. We consider two wc4DVAR formulations in this thesis, the model error formulation and state estimation formulation. The 4DVAR objective function is traditionally solved using gradient-based iterative methods. The principle method used in Numerical Weather Prediction today is the Gauss-Newton approach. This method introduces a linearised `inner-loop' objective function, which upon convergence, updates the solution of the non-linear `outer-loop' objective function. This requires many evaluations of the objective function and its gradient, which emphasises the importance of the Hessian. The eigenvalues and eigenvectors of the Hessian provide insight into the degree of convexity of the objective function, while also indicating the difficulty one may encounter while iterative solving 4DVAR. The condition number of the Hessian is an appropriate measure for the sensitivity of the problem to input data. The condition number can also indicate the rate of convergence and solution accuracy of the minimisation algorithm. This thesis investigates the sensitivity of the solution process minimising both wc4DVAR objective functions to the internal assimilation parameters composing the problem. We gain insight into these sensitivities by bounding the condition number of the Hessians of both objective functions. We also precondition the model error objective function and show improved convergence. We show that both formulations' sensitivities are related to error variance balance, assimilation window length and correlation length-scales using the bounds. We further demonstrate this through numerical experiments on the condition number and data assimilation experiments using linear and non-linear chaotic toy models.

Balanced Ocean-Data Assimilation near the Equator

The question is addressed whether using unbalanced updates in ocean-data assimilation schemes for seasonal forecasting systems can result in a relatively poor simulation of zonal currents. An assimilation scheme, where temperature observations are used for updating only the density field, is compared to a scheme where updates of density field and zonal velocities are related by geostrophic balance. This is done for an equatorial linear shallow-water model. It is found that equatorial zonal velocities can be detoriated if velocity is not updated in the assimilation procedure. Adding balanced updates to the zonal velocity is shown to be a simple remedy for the shallow-water model. Next, optimal interpolation (OI) schemes with balanced updates of the zonal velocity are implemented in two ocean general circulation models. First tests indicate a beneficial impact on equatorial upper-ocean zonal currents.

Reducing defects through organizational learning within a Housing Association environment

Housing Associations (HAs) contribute circa 20% of the UK’s housing supply. HAs are however under increasing pressure as a result of funding cuts and rent reductions. Due to the increased pressure, a number of processes are currently being reviewed by HAs, especially how they manage and learn from defects. Learning from defects is considered a useful approach to achieving defect reduction within the UK housebuilding industry. This paper contributes to our understanding of how HAs learn from defects by undertaking an initial round table discussion with key HA stakeholders as part of an ongoing collaborative research project with the National House Building Council (NHBC) to better understand how house builders and HAs learn from defects to reduce their prevalence. The initial discussion shows that defect information runs through a number of groups, both internal and external of a HA during both the defects management process and organizational learning (OL) process. Furthermore, HAs are reliant on capturing and recording defect data as the foundation for the OL process. During the OL process defect data analysis is the primary enabler to recognizing a need for a change to organizational routines. When a need for change has been recognized, new options are typically pursued to design out defects via updates to a HAs Employer’s Requirements. Proposed solutions are selected by a review board and committed to organizational routine. After implementing a change, both structured and unstructured feedback is sought to establish the change’s success. The findings from the HA discussion demonstrates that OL can achieve defect reduction within the house building sector in the UK. The paper concludes by outlining a potential ‘learning from defects model’ for the housebuilding industry as well as describing future work.

Heterogeneous tensor decomposition for clustering via manifold optimization

Tensor clustering is an important tool that exploits intrinsically rich structures in real-world multiarray or Tensor datasets. Often in dealing with those datasets, standard practice is to use subspace clustering that is based on vectorizing multiarray data. However, vectorization of tensorial data does not exploit complete structure information. In this paper, we propose a subspace clustering algorithm without adopting any vectorization process. Our approach is based on a novel heterogeneous Tucker decomposition model taking into account cluster membership information. We propose a new clustering algorithm that alternates between different modes of the proposed heterogeneous tensor model. All but the last mode have closed-form updates. Updating the last mode reduces to optimizing over the multinomial manifold for which we investigate second order Riemannian geometry and propose a trust-region algorithm. Numerical experiments show that our proposed algorithm compete effectively with state-of-the-art clustering algorithms that are based on tensor factorization.