Search

All major weather and climate applications are currently developed using languages such as Fortran or C++. This is typical in the domain of high performance computing (HPC), where efficient execution is an important concern. Unfortunately, this approach leads to implementations that intermix optimizations for specific hardware architectures with the high-level numerical methods that are typical for the domain. This leads to code that is verbose, difficult to extend and maintain, and difficult to port to different hardware architectures. Here, we propose a different strategy based on GT4Py (GridTools for Python). GT4Py is a Python framework to write weather and climate applications that includes a high-level embedded domain specific language (DSL) to write stencil computations. The toolchain integrated in GT4Py enables automatic code-generation,to obtain the performance of state-of-the-art C++ and CUDA implementations. The separation of concerns between the mathematical definitions and the actual implementations allows for performance portability of the computations on a wide range of computing architectures, while being embedded in Python allows easy access to the tools of the Python ecosystem to enhance the productivity of the scientists and facilitate integration in complex workflows. Here, the initial release of GT4Py is described, providing an overview of the current state of the framework and performance results showing how GT4Py can outperform pure Python implementations by orders of magnitude., Comment: 12 pages

OpenQ$^\star$D code has been used by the RC$^\star$ collaboration for the generation of fully dynamical QCD+QED gauge configurations with C$^\star$ boundary conditions. In this talk, optimization of solvers provided with the openQ$^\star$D package relevant for porting the code on GPU-accelerated supercomputing platforms is discussed. We present the analysis of the current implementations of the GCR solver preconditioned with Schwarz alternating procedure for ill-conditioned Dirac-operators. With the goal of enabling support for GPUs from various vendors, a novel method of adaptive CPU/GPU-hybrid implementation is proposed.

Materials Cloud is a platform designed to enable open and seamless sharing of resources for computational science, driven by applications in materials modelling. It hosts 1) archival and dissemination services for raw and curated data, together with their provenance graph, 2) modelling services and virtual machines, 3) tools for data analytics, and pre-/post-processing, and 4) educational materials. Data is citable and archived persistently, providing a comprehensive embodiment of the FAIR principles that extends to computational workflows. Materials Cloud leverages the AiiDA framework to record the provenance of entire simulation pipelines (calculations performed, codes used, data generated) in the form of graphs that allow to retrace and reproduce any computed result. When an AiiDA database is shared on Materials Cloud, peers can browse the interconnected record of simulations, download individual files or the full database, and start their research from the results of the original authors. The infrastructure is agnostic to the specific simulation codes used and can support diverse applications in computational science that transcend its initial materials domain., Comment: 19 pages, 8 figures

The dynamical cluster approximation (DCA) is a quantum cluster extension to the single-site dynamical mean-field theory that incorporates spatially nonlocal dynamic correlations systematically and nonperturbatively. The DCA$^+$ algorithm addresses the cluster shape dependence of the DCA and improves the convergence with cluster size by introducing a lattice self-energy with continuous momentum dependence. However, we show that the DCA$^+$ algorithm is plagued by a fundamental problem when its self-consistency equations are formulated using the bare Green's function of the cluster. This problem is most severe in the strongly correlated regime at low doping, where the DCA$^+$ self-energy becomes overly metallic and local, and persists to cluster sizes where the standard DCA has long converged. In view of the failure of the DCA$^+$ algorithm, we propose to complement DCA simulations with a post-interpolation procedure for single-particle and two-particle correlation functions to preserve continuous momentum dependence and the associated benefits in the DCA. We demonstrate the effectiveness of this practical approach with results for the half-filled and hole-doped two-dimensional Hubbard model., Comment: 11 pages, 10 figures

We present the first open release of the DCA++ project, a high-performance research software framework to solve quantum many-body problems with cutting edge quantum cluster algorithms. DCA++ implements the dynamical cluster approximation (DCA) and its DCA$^+$ extension with a continuous self-energy. The algorithms capture nonlocal correlations in strongly correlated electron systems, thereby giving insight into high-$T_c$ superconductivity. The code's scalability allows efficient usage of systems at all scales, from workstations to leadership computers. With regard to the increasing heterogeneity of modern computing machines, DCA++ provides portable performance on conventional and emerging new architectures, such as hybrid CPU-GPU, sustaining multiple petaflops on ORNL's Titan and CSCS' Piz Daint supercomputers. Moreover, we show how sustainable and scalable development of the code base has been achieved by adopting standard techniques of the software industry. These include employing a distributed version control system, applying test-driven development and following continuous integration., Comment: 32 pages, 17 figures, 1 table, 3 code listings; submitted to Comput. Phys. Commun

The GW approximation is a well-known method to improve electronic structure predictions calculated within density functional theory. In this work, we have implemented a computationally efficient GW approach that calculates central properties within the Matsubara-time domain using the modified version of Elk, the full-potential linearized augmented plane wave (FP-LAPW) package. Continuous-pole expansion (CPE), a recently proposed analytic continuation method, has been incorporated and compared to the widely used Pade approximation. Full crystal symmetry has been employed for computational speedup. We have applied our approach to 18 well-studied semiconductors/insulators that cover a wide range of band gaps computed at the levels of single-shot G0W0, partially self-consistent GW0, and fully self-consistent GW (scGW). Our calculations show that G0W0 leads to band gaps that agree well with experiment for the case of simple s-p electron systems, whereas scGW is required for improving the band gaps in 3-d electron systems. In addition, GW0 almost always predicts larger band gap values compared to scGW, likely due to the substantial underestimation of screening effects. Both the CPE method and Pade approximation lead to similar band gaps for most systems except strontium titantate, suggesting further investigation into the latter approximation is necessary for strongly correlated systems. Our computed band gaps serve as important benchmarks for the accuracy of the Matsubara-time GW approach., Comment: 4 tables, 3 figures

In October 2013 CSCS installed the first hybrid Cray XC-30 system, dubbed Piz Daint. This system features the power management database (PMDB), that was recently introduced by Cray to collect detailed power consumption information in a non-intrusive manner. Power measurements are taken on each node, with additional measurements for the Aries network and blowers, and recorded in a database. This enables fine-grained reporting of power consumption that is not possible with external power meters, and is useful to both application developers and facility operators. This paper will show how benchmarks of representative applications at CSCS were used to validate the PMDB on Piz Daint. Furthermore we will elaborate, with the well-known HPL benchmark serving as prototypical application, on how the PMDB streamlines the tuning for optimal power efficiency in production, which lead to Piz Daint being recognised as the most energy efficient petascale supercomputer presently in operation., Comment: This paper was presented at the 2014 Cray User Group (CUG) user meeting in Lugano, Switzerland,First Experiences With Validating and Using the Cray Power Management Database Tool, Gilles Fourestey and Ben Cumming and Ladina Gilly, Proceedings of the CUG meeting, 2014

We have investigated the group 14 nitrides (M$_3$N$_4$) in the spinel phase ($\gamma$-M$_3$N$_4$ with M= C, Si, Ge and Sn) and $\beta$ phase ($\beta$-M$_3$N$_4$ with M= Si, Ge and Sn) using density functional theory with the local density approximation and the GW approximation. The Kohn-Sham energies of these systems have been first calculated within the framework of full-potential linearized augmented plane waves and then corrected using single-shot G$_0$W$_0$ calculations, which we have implemented in the modified version of the Elk full-potential LAPW code. Direct band gaps at the $\Gamma$ point have been found for spinel-type nitrides $\gamma$-M$_3$N$_4$ with M= Si, Ge and Sn. The corresponding GW-corrected band gaps agree with experiment. We have also found that the GW calculations with and without the plasmon-pole approximation give very similar results, even when the system contains semi-core $d$ electrons. These spinel-type nitrides are novel materials for potential optoelectronic applications because of their direct and tunable band gaps., Comment: 8 pages, 5 figures

The dynamical cluster approximation (DCA) is a systematic extension beyond the single site approximation in dynamical mean field theory (DMFT), to include spatially non-local correlations in quantum many-body simulations of strongly correlated systems. We extend the DCA with a continuous lattice self-energy in oder to achieve better convergence with cluster size. The new method, which we call DCA$^+$, cures the cluster shape dependence problems of the DCA, without suffering from causality violations of previous attempts to interpolate the cluster self-energy. A practical approach based on standard inference techniques is given to deduce the continuous lattice self-energy from an interpolated cluster self-energy. We study the pseudogap region of a hole-doped two-dimensional Hubbard model and find that in the DCA$^+$ algorithm, the self-energy and pseudo-gap temperature $T^*$ converge monotonously with cluster size. Introduction of a continuous lattice self-energy eliminates artificial long-rage correlations and thus significantly reduces the sign problem of the quantum Monte Carlo cluster solver in the DCA$^+$ algorithm compared to the normal DCA. Simulations with much larger cluster sizes thus become feasible, which, along with the improved convergence in cluster size, raises hope that precise extrapolations to the exact infinite cluster size limit can be reached for other physical quantities as well.

The adoption of hybrid GPU-CPU nodes in traditional supercomputing platforms opens acceleration opportunities for electronic structure calculations in materials science and chemistry applications, where medium sized Hermitian generalized eigenvalue problems must be solved many times. The small size of the problems limits the scalability on a distributed memory system, hence they can benefit from the massive computational performance concentrated on a single node, hybrid GPU-CPU system. However, new algorithms that efficiently exploit heterogeneity and massive parallelism of not just GPUs, but of multi/many-core CPUs as well are required. Addressing these demands, we implemented a novel Hermitian general eigensolver algorithm. This algorithm is based on a standard eigenvalue solver, and existing algorithms can be used. The resulting eigensolvers are state-of-the-art in HPC, significantly outperforming existing libraries. We analyze their performance impact on applications of interest, when different fractions of eigenvectors are needed by the host electronic structure code.

By considering the appropriate finite-size effect, we explain the connection between Monte Carlo simulations of two-dimensional anisotropic Heisenberg antiferromagnet in a field and the early renormalization group calculation for the bicritical point in $2+\epsilon$ dimensions. We found that the long length scale physics of the Monte Carlo simulations is indeed captured by the anisotropic nonlinear $\sigma$ model. Our Monte Carlo data and analysis confirm that the bicritical point in two dimensions is Heisenberg-like and occurs at T=0, therefore the uncertainty in the phase diagram of this model is removed., Comment: 10 pages, 11 figures

We have studied magnetic ordering in ferro/antiferromagnetic (F/AF) bilayers using Monte Carlo simulations of classical Heisenberg spins. For both flat and stepped interfaces we observed order in the AF above the Neel temperature, with the AF spins aligning collinearly with the F moments. In the case of the stepped interface there is a transition from collinear to perpendicular alignment of the F and AF spins at a lower temperature., Comment: Latex, 2 pages, 2 figures (proceedings of ICM2003)

Monte Carlo simulations have been used to study phase transitions on coupled anisotropic ferro/antiferromagnetic (FM/AFM) films of classical Heisenberg spins. We consider films of different thicknesses, with fully compensated exchange across the FM/AFM interface. We find indications of a phase transition on each film, occuring at different temperatures. It appears that both transition temperatures depend on the film thickness., Comment: Revtex, 4 pages, 4 figures

Monte Carlo simulations have been used to study magnetic ordering in coupled anisotropic ferro/antiferromagnetic (FM/AFM) films of classical Heisenberg spins. We consider films with flat interfaces that are fully uncompensated as well as rough interfaces that are compensated on average. For both types of interfaces above the ``N\'eel temperature'' we observed order in the AFM with the AFM spins aligning collinearly with the FM moments. In the case of rough interfaces there is a transition from collinear to perpendicular alignment of the FM and AFM spins at a lower temperature., Comment: Revtex, 4 pages, 4 figures