Weak Scale Research Articles

From GPU to CPU (and Beyond): Extending Hardware Support in GPUSPH Through a SYCL‐Inspired Interface

ABSTRACTWhile most software is originally designed for serial or parallel execution on CPU, and porting to GPU comes later in its development, GPUSPH was designed from the ground up to run on GPUs using CUDA. Making it accessible to a wider audience by introducing support for other computational hardware, and in particular CPUs, poses challenges that are complementary to the ones normally faced when porting CPU code to GPU. We present the approach we have adopted to support CPUs as computational devices in GPUSPH with minimal code changes and low developer effort. Detailed benchmarks illustrating the performance of the implementation and its scalability across multiple cores in both single‐socket and NUMA configurations show good strong and weak scaling.

Decoding the Gaugino Code Naturally at High-Lumi LHCs

Natural supersymmetry with light higgsinos is most favored to emerge from the string landscape, since the volume of a scan parameter space shrinks to tiny volumes for electroweak unnatural models. Rather general arguments favor a landscape selection of soft SUSY breaking terms tilted to large values, but they are tempered by the atomic principle that the derived value of the weak scale in each pocket universe lies not too far from its measured value in our universe. But, that leaves (at least) three different paradigms for gaugino masses in natural SUSY models: unified (as in nonuniversal Higgs models), anomaly mediation form (as in natural AMSB), and mirage mediation form (with comparable moduli- and anomaly-mediated contributions). We perform landscape scans for each of these, and we show that they populate different, but overlapping, positions in m(ℓℓ¯) and m(wino) space. The first of these may be directly measurable at high-lumi LHC via the soft opposite-sign dilepton plus jets plus E/T signature arising from higgsino pair production, while the second of these could be extracted from direct wino pair production, leading to same-sign diboson production.

Bringing the Peccei-Quinn mechanism down to Earth

It is conventionally assumed that the physics underlying the Peccei-Quinn (PQ) mechanism for addressing the strong CP problem is at very high energies, orders of magnitude above the weak scale. However, this may not be the case in general and the associated PQ boson ϕ, besides the signature state, i.e., the ultralight axion a, may emerge well below the weak scale. We consider this possibility and examine some of the conditions for its viability. The example model proposed here may also provide the requisite Standard Model Higgs mass parameter, without invoking new scalars above the giga-electron-volt (GeV) scale. The corresponding parameter space can maintain against quantum corrections. This scenario, depending on the choice of parameters, can potentially be constrained by flavor data. We point out that the current mild excess in B+→K+νν¯, reported by the Belle II experiment, could in principle be explained in this setup as B+→K+ϕ and B+→K+a, with both ϕ and a escaping the detector as missing energy. For a sufficiently heavy PQ boson, in the GeV regime (as the preferred explanation of the Belle II excess), one can separate these two contributions, due to the difference in K+ momenta. In this case, the axion may also affect lighter meson, e.g., kaon, decays while ϕ would not be a kinematically allowed final state. Published by the American Physical Society 2024

High-speed turbulent flows towards the exascale: STREAmS-2 porting and performance

Exascale High Performance Computing (HPC) represents a tremendous opportunity to push the boundaries of Computational Fluid Dynamics (CFD), but despite the consolidated trend towards the use of Graphics Processing Units (GPUs), programmability is still an issue. STREAmS-2 (Bernardini et al. Comput. Phys. Commun. 285 (2023) 108644) is a compressible solver for canonical wall-bounded turbulent flows capable of harvesting the potential of NVIDIA GPUs. Here we extend the already available CUDA Fortran backend with a novel HIP backend targeting AMD GPU architectures. The main implementation strategies are discussed along with a novel Python tool that can generate the HIP and CPU code versions allowing developers to focus their attention only on the CUDA Fortran backend. Single GPU performance is analysed focusing on NVIDIA A100 and AMD MI250x cards which are currently at the core of several HPC clusters. The gap between peak GPU performance and STREAmS-2 performance is found to be generally smaller for NVIDIA cards. Roofline analysis allows tracing this behavior to unexpectedly different computational intensities of the same kernel using the two cards. Additional single-GPU comparisons are performed to assess the impact of grid size, number of parallelized loops, thread masking and thread divergence. Parallel performance is measured on the two largest EuroHPC pre-exascale systems, LUMI (AMD GPUs) and Leonardo (NVIDIA GPUs). Strong scalability reveals more than 80% efficiency up to 16 nodes for Leonardo and up to 32 for LUMI. Weak scalability shows an impressive efficiency of over 95% up to the maximum number of nodes tested (256 for LUMI and 512 for Leonardo). This analysis shows that STREAmS-2 is the perfect candidate to fully exploit the power of current pre-exascale HPC systems in Europe, allowing users to simulate flows with over a trillion mesh points, thus reducing the gap between the Reynolds numbers achievable in high-fidelity simulations and those of real engineering applications.

Fast inspirals and the treatment of orbital resonances

Abstract Extreme mass ratio inspirals (EMRIs), where a compact object orbits a massive black hole, are a key source of gravitational waves for the future Laser Interferometer Space Antenna (LISA). Due to their small mass ratio, ε ~ 10-4 -- 10-7 , the binary evolves slowly and EMRI signals will be in-band for years. Additionally, astrophysical EMRIs are expected to have complex dynamics featuring both spin-precession and eccentricity. A standard approach to modelling these inspirals is via the method of osculating geodesics (OG) which we employ along with a toy model for the gravitational self-force. Using this method requires resolving tens of thousands radial and polar orbital librations over the long duration of the signal which makes the inspiral trajectory expensive to compute.In this work we accelerate these calculations by employing Near-Identity (averaging) Transformations. 
However, this averaging technique breaks down at orbital resonances where the radial and polar frequencies are an integer ratio of each other. Thus, we switch to a partial averaging transformation in the vicinity of the resonance where the dynamics are characterised by the slow evolution of the so-called ``resonant phase". Additionally, we develop an optimal switching criterion to minimise the computation time while maximising accuracy. We find the error in the waveform phase is improved from O(ε-1/2) in the fully averaged scheme to O(ε4/7) in the switching scheme. At the same time, this scheme improves the scaling of the computation time from being inversely proportional to ε using OG, to a very weak scaling with ε. This results in a speed-up of at least two orders of magnitude for LISA EMRIs with room for further optimisation.

Sustainable rewritable paper based on photoresponsive tungsten oxide quantum dots for anti-counterfeiting and waterproofing

Ink-free printing using photochromic paper has emerged as a promising technology for information recording, transmission, and secure interaction applications. Tungsten oxide quantum dots (WO3 QDs), with their rapid UV response, excellent photochemical stability, and outstanding fatigue resistance, exhibit significant potential as an ink-free medium in the construction of sustainable rewritable paper. However, challenges such as low binding force, poor stability, and weak scalability limit the development of reversible photochromic paper based on WO3. In this study, a novel sustainable rewritable paper based on a scalable strategy of covalent modification and hydrophobic layer protection has been reported. Specifically, cellulose papers modified with silane coupling agent were directly grafted with amino-modified WO3 QDs by covalent bonding through the addition reaction between epoxy and amine groups. Subsequently, the sustainable rewritable paper was obtained by the deposition of hydrophobic layer of polydimethylsiloxane (PDMS) via a dip-coating method. The rewritable paper exhibits exceptional properties, including high color contrast, extended color-retention (over 21 days), good rewritability (20 cycles) and excellent resistance to water or stains. Based on ink-free writing technology, the photo-responsive rewritable papers were utilized for anti-counterfeiting and encryption or decryption, demonstrating their great potential for information recording, storage, and security protection.

Resolving the ultracollinear paradox with effective field theory

Naive intuition about scale decoupling breaks down in the presence of fermion masses. Kinematic enhancements can greatly extend the range where one needs to keep a finite mass in calculations to obtain a correct result at even the O(1) level. Treating a light fermion as massive though, leads to a known but somewhat obscure paradox, a seeming leading-order sensitivity to an arbitrarily small mass. We show how a proper formulation in effective field theory not only resolves the physical conundrum but also answers the very practical question of when fermion masses are required. This has important implications for the development of shower Monte Carlo event generators above the weak scale. Published by the American Physical Society 2024

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count.Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 20483 on a single node to 327683 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 20483 where communication is entirely intra-node, and for 327683, where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 327683 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

A multiverse outside of the swampland

A multiverse can arise from landscapes without de Sitter minima. It can be populated during a period of eternal inflation without trans-Planckian field excursions and without flat potentials. This multiverse can explain the values of the cosmological constant and of the weak scale. In the process of proving these statements, we derive a few simple, but counterintuitive results. We show that it is easy to write models of eternal inflation compatible with the distance and refined de Sitter conjectures. Secondly, tunneling transitions that move fields from a lower-energy vacuum to a higher-energy vacuum and generate baby universes are possible, and occur during eternal inflation. Finally, we relax the assumption of no de Sitter minima and show that this more standard multiverse can be populated by Coleman-de Luccia transitions in about 100 e-folds of inflation. Published by the American Physical Society 2024

Local Second Order Mo̷ller-Plesset Theory with a Single Threshold Using Orthogonal Virtual Orbitals: A Distributed Memory Implementation.

In order to alleviate the computational burden associated with superlinear compute scalings with molecular size in electron correlation methods, researchers have developed local correlation methods that wisely treat relatively small contributions as zeros but still yield accurate energy approximation. Such local correlation techniques can also be combined with parallel computing resources to obtain further efficiency and scalability. This work focuses on the distributed memory parallel implementation of a local correlation method for second order Mo̷ller-Plesset (MP2) theory. This method also only has a single threshold to control the dropping of terms and accuracy of different computing kernels in the algorithm. The process partitioning strategy and distributed parallel implementation with the message passing interface (MPI) are discussed. In particular, the algorithm relies on a fixed sparsity pattern matrix multiplication and a corresponding distributed conjugate gradient solver, which exhibits almost linear scaling in both strong and weak scaling analyses. Numerical experiments on a range of molecules, including linear chains and molecules with 2 and 3-dimensional characters, are reported. For example, with only 32 MPI ranks, this MP2 implementation can calculate the correlation energy of vancomycin in def2-TZVP basis within 0.003% accuracy (10-6.5 threshold) in half an hour, where the same problem is unfeasible to solve with sequential or pure shared memory implementations.

Superior ablation resistance of plasma sprayed SiC based coating by structural optimized powders

Pristine SiC coating prepared by plasma spraying is difficult to achieve good ablation resistance at high temperature due to its weak SiO2 scale. Besides incorporation by ultra-high temperature ceramics, the modification of the SiC coating from its feedstock powder is necessary. In the current work, the mixed powders of SiC, Si and C, where the SiC was encapsulated and then carbonized by polydopamine, denoted as SiC@C–Si, were employed as the feedstocks, compared with the mixed case, i.e. SiC–Si–C. The SiC@C–Si coating exhibited superior anti-ablation properties, with a linear ablation rate of 0.62 μm/s, 57 % lower than the SiC–Si–C coating. This is attributed to the fact that nano-carbon formed by carbonization of polydopamine is more prone to reacting with Si than micron-sized C powder, thereby forming a dense microstructure after ablation. Besides, in-situ formed SiC by the depletion reaction increases the flow resistance and decelerates the failure of SiO2 film, thus improving the ablation resistance.

Massively parallel axisymmetric fluid model for streamer discharges

A highly parallelizable fluid plasma simulation tool based upon the first-order drift-diffusion equations is discussed. Atmospheric pressure plasmas have densities and gradients that require small element sizes in order to accurately simulate the plasm resulting in computational meshes on the order of millions to tens of millions of elements for realistic size plasma reactors. To enable simulations of this nature, parallel computing is required and must be optimized for the particular problem. Here, a finite-volume, electrostatic drift-diffusion implementation for low-temperature plasma is discussed. The implementation is built upon the Message Passing Interface (MPI) library in C++ using Object Oriented Programming. The underlying numerical method is outlined in detail and benchmarked against simple streamer formation from other streamer codes. Electron densities, electric field, and propagation speeds are compared with the reference case and show good agreement. Convergence studies are also performed showing a minimal space step of approximately 4 μm required to reduce relative error to below 1% during early streamer simulation times and even finer space steps are required for longer times. Additionally, strong and weak scaling of the implementation are studied and demonstrate the excellent performance behavior of the implementation up to 100 million elements on 1024 processors. Finally, different advection schemes are compared for the simple streamer problem to analyze the influence of numerical diffusion on the resulting quantities of interest.

Toward an efficient second‐order method for computing the surface gravitational potential on spherical‐polar meshes

AbstractAstrophysical accretion discs that carry a significant mass compared with their central object are subject to the effect of self‐gravity. In the context of circumstellar discs, this can, for instance, cause fragmentation of the disc gas, and—under suitable conditions—lead to the direct formation of gas‐giant planets. If one wants to study these phenomena, the disc's gravitational potential needs to be obtained by solving the Poisson equation. This requires to specify suitable boundary conditions. In the case of a spherical‐polar computational mesh, a standard multipole expansion for obtaining boundary values is not practicable. We hence compare two alternative methods for overcoming this limitation. The first method is based on a known Green's function expansion (termed “CCGF”) of the potential, while the second (termed “James' method”) uses a surface screening mass approach with a suitable discrete Green's function. We demonstrate second‐order convergence for both methods and test the weak scaling behavior when using thousands of computational cores. Overall, James' method is found superior owing to its favorable algorithmic complexity of compared with the scaling of the CCGF method.

Thermoviscous dissipation of nonlinear acoustic waves in channels with wavy walls.

We derive a nonlinear acoustic wave propagation model for analyzing the thermoviscous dissipation of nonlinear acoustic waves in narrow pores with wavy walls using the boundary layer theory. As a nonlinear acoustic wave propagates in a pore, the wave-steepening effect competes with the bulk dissipation, as well as the thermoviscous heat transfer and shear from the pore walls. Due to thermoviscous dissipation, the wave thickness increases beyond the weak shock thickness scale. Using the weak shock thickness scale, we obtain dimensionless linear and nonlinear model wave equations governing the shock-wall interactions. We also perform two-dimensional shock-resolved direct numerical simulation of the wave propagation inside the pores and compare the results with model equations. The direct numerical simulation and model calculations show that, for flat walls and shock strength parameter ϵ, the dimensional wall heat-flux and shear scale as ϵ. For wavy walls, the scaling becomes ϵ3/2-n(k) where k is the wall-waviness wavenumber and the exponent n increases from 0.5 for k = 0 to n(k)≈0.65 for k = 10, n(k)≈0.75 for k = 20, and n(k)≈0.85 for k = 40. Furthermore, we show that both the dimensionless scaled wall shear and wall heat-flux decrease with increasing k.

An approximate block factorization preconditioner for mixed-dimensional beam-solid interaction

This paper presents a scalable approximate block factorization preconditioner for mixed-dimensional models in beam-solid interaction and their application in engineering. In particular, it studies the linear systems arising from a regularized mortar-type approach for embedding geometrically exact beams into solid continua. Due to the lack of block diagonal dominance of the arising 2 × 2 block system, an approximate Block-LU preconditioner is used. It exploits the sparsity structure of the beam sub-block to construct a sparse approximate inverse, which is then not only used to explicitly form an approximation of the Schur complement, but also acts as a smoother within the prediction and correction step of the arising Block-LU preconditioner. The Schur complement equation is tackled with an algebraic multigrid method. Although, for now, the beam sub-block is tackled by a one-level method only, the multi-level nature of the computationally demanding Schur equation delivers a scalable preconditioner in practice. In numerical test cases, the influence of different algorithmic parameters on the quality of the sparse approximate inverse is studied and the weak scaling behavior of the proposed preconditioner on up to 1000 MPI ranks is demonstrated. In addition, the robustness of the proposed method regarding material parameters and geometric properties is shown, before the preconditioner is finally applied for the analysis of steel-reinforced concrete structures in civil engineering.

Composite dark matter and neutrino masses from a light hidden sector

We study a class of models in which the particle that constitutes dark matter arises as a composite state of a strongly coupled hidden sector. The hidden sector interacts with the Standard Model through the neutrino portal, allowing the relic abundance of dark matter to be set by annihilation into final states containing neutrinos. The coupling to the hidden sector also leads to the generation of neutrino masses through the inverse seesaw mechanism, with composite hidden sector states playing the role of the singlet neutrinos. We focus on the scenario in which the hidden sector is conformal in the ultraviolet, and the compositeness scale lies at or below the weak scale. We construct a holographic realization of this framework based on the Randall-Sundrum setup and explore the implications for experiments. We determine the current constraints on this scenario from direct and indirect detection, lepton flavor violation and collider experiments and explore the reach of future searches. We show that in the near future, direct detection experiments and searches for μ → e conversion will be able to probe new parameter space. At colliders, dark matter can be produced in association with composite singlet neutrinos via Drell Yan processes or in weak decays of hadrons. We show that current searches at the Large Hadron Collider have only limited sensitivity to this new production channel and we comment on how the reconstruction of the singlet neutrinos can potentially expand the reach.

Electroweak evolution equations and isospin conservation

In processes taking place at energies much higher than the weak scale, electroweak corrections can be taken into account by using electroweak evolution equations, that are analogous to the DGLAP equations in QCD. We show that weak isospin conservation in these equations imposes to modify the expressions of the splitting functions commonly used in the literature. These modifications have a profound impact on the parton distribution functions.

Weak time-scale separation at the onset of oscillatory magnetoconvection in rapidly rotating fluids

Convective instabilities are one of the integral parts of the dynamics of flows driven by thermal buoyancy. Naturally, physical phenomena exhibit a wide disparity in the length and timescales of the field variables in numerical simulations and experimental observations. Such variations are not represented in the traditional normal mode stability analysis attempting to understand the onset of convection. This study attempts to incorporate different time constants for different perturbation variables in the linear stability analysis with the help of a Taylor series expansion. The infinite horizontal layer model is chosen for simplicity. Apart from the classical Rayleigh-Bénard system, additional physical effects such as background rotation and magnetic field have been considered with plausible implications for geophysical flow applications. The time scale separation is implemented by considering a slight change in the frequency of temperature perturbation compared to that for other physical quantities. Both analytical and numerical methods have been utilised for the investigation. The threshold buoyancy force is reduced when the temperature perturbation has a smaller frequency than the frequencies of other variables. Besides that, the onset wavenumber and frequency of the oscillatory modes are modified due to weak scale separation from the onset characteristics of the reference case. In particular, enhanced frequency of temperature perturbations leads to smaller-scaled magnetically controlled convective rolls and larger-scaled viscously controlled instabilities at the onset. A robust dependence of the onset characteristics with the parameter quantifying the timescale separation is obtained. Additionally, two transitions in the convective onset modes with scale separation have been identified.

Stau pairs from natural SUSY at high luminosity LHC

Natural supersymmetry (SUSY) with light Higgsinos is perhaps the most plausible of all weak scale SUSY models while a variety of motivations point to (right) tau sleptons as the lightest of all the sleptons. We examine a SUSY model line with rather light right staus embedded within natural SUSY. For light τ˜1 of a few hundred GeV, the decays τ˜1→τχ˜1,20 and ντχ˜1− occur at comparable rates where the (Higgsino-like) χ˜1± and χ˜20 release only small visible energy: in this case, the expected τ+τ−+ET signature is diminished from the usual expectations due to the presence of the nearly invisible decay mode τ˜1→ντχ˜1−. However, once mτ˜1≳m(bino), decays to binos such as τ˜1→τχ˜30 open up where χ˜30 decays to Higgsinos plus W±, Z0, and h at comparable rates. For these heavier staus, the stau pair production gives rise to diboson+ET events, which may contain 0, 1, or 2 additional hard τ leptons. From these considerations, we examine the potential for future discovery of tau-slepton pair production at a high-luminosity LHC. While we do not find a 5σ HL-LHC discovery reach for 3000 fb−1, we do find a 95% CL exclusion reach, ranging between mτ˜1: 100–450 GeV for mχ˜10∼100 GeV. This latter reach disappears for mχ˜10≳200 GeV. Published by the American Physical Society 2024

Asynchronous global–local non-invasive coupling for nonlinear monotone patches: Application to plasticity problems

This article presents the asynchronous global–local non-invasive coupling in the case of nonlinear monotone patches as encountered when inserting elastoplastic components in a global linear elastic model. The convergence of the method is theoretically established in a general framework using the paracontractions technique and illustrated in academic examples with a weak scalability study considering the case of balanced and unbalanced patches.