arXiv subjects

Patrick Sanan

Publications and source records attributed to Patrick Sanan.

9 recordsLinked to original sources

arXiv2022

Retention of water in terrestrial magma oceans and carbon-rich early atmospheres

Massive steam and CO$_2$ atmospheres have been proposed for magma ocean outgassing of Earth and terrestrial planets. Yet formation of such atmospheres depends on volatile exchange with the molten interior, governed by volatile solubilities and redox reactions. We determine the evolution of magma ocean--atmosphere systems for a range of oxygen fugacities, C/H ratios and hydrogen budgets that include redox reactions for hydrogen (H$_2$--H$_2$O), carbon (CO--CO$_2$), methane (CH$_4$), and solubility laws for H$_2$O and CO$_2$. We find that small initial budgets of hydrogen, high C/H ratios, and oxidizing conditions, suppress outgassing of hydrogen until the late stage of magma ocean crystallization. Hence early atmospheres in equilibrium with magma oceans are dominantly carbon-rich, and specifically CO-rich except at the most oxidizing conditions. The high solubility of H$_2$O limits its outgassing to melt fractions below $\sim$30\%, the fraction at which the mantle transitions from vigorous to sluggish convection with melt percolation. Sluggish melt percolation could enable a surface lid to form, trapping water in the interior and thereby maintaining a carbon-rich atmosphere (equilibrium crystallization). Alternatively, efficient crystal settling could maintain a molten surface, promoting a transition to a water-rich atmosphere (fractional crystallization). However, additional processes, including melt trapping and H dissolution in crystallizing minerals, further conspire to limit the extent of H outgassing, even for fractional crystallization. Hence, much of the water delivered to planets during their accretion can be safely harbored in their interiors during the magma ocean stage, particularly at oxidizing conditions.

astro-ph.EP↗

arXiv2022

The PETSc Community Is the Infrastructure

The communities who develop and support open source scientific software packages are crucial to the utility and success of such packages. Moreover, these communities form an important part of the human infrastructure that enables scientific progress. This paper discusses aspects of the PETSc (Portable Extensible Toolkit for Scientific Computation) community, its organization, and technical approaches that enable community members to help each other efficiently.

cs.SE↗

arXiv2021

Understanding performance variability in standard and pipelined parallel Krylov solvers

In this work, we collect data from runs of Krylov subspace methods and pipelined Krylov algorithms in an effort to understand and model the impact of machine noise and other sources of variability on performance. We find large variability of Krylov iterations between compute nodes for standard methods that is reduced in pipelined algorithms, directly supporting conjecture, as well as large variation between statistical distributions of runtimes across iterations. Based on these results, we improve upon a previously introduced nondeterministic performance model by allowing iterations to fluctuate over time. We present our data from runs of various Krylov algorithms across multiple platforms as well as our updated non-stationary model that provides good agreement with observations. We also suggest how it can be used as a predictive tool.

cs.MS↗

arXiv2021

Vertically resolved magma ocean-protoatmosphere evolution: H$_2$, H$_2$O, CO$_2$, CH$_4$, CO, O$_2$, and N$_2$ as primary absorbers

The earliest atmospheres of rocky planets originate from extensive volatile release during magma ocean epochs that occur during assembly of the planet. These establish the initial distribution of the major volatile elements between different chemical reservoirs that subsequently evolve via geological cycles. Current theoretical techniques are limited in exploring the anticipated range of compositional and thermal scenarios of early planetary evolution, even though these are of prime importance to aid astronomical inferences on the environmental context and geological history of extrasolar planets. Here, we present a coupled numerical framework that links an evolutionary, vertically-resolved model of the planetary silicate mantle with a radiative-convective model of the atmosphere. Using this method we investigate the early evolution of idealized Earth-sized rocky planets with end-member, clear-sky atmospheres dominated by either H$_2$, H$_2$O, CO$_2$, CH$_4$, CO, O$_2$, or N$_2$. We find central metrics of early planetary evolution, such as energy gradient, sequence of mantle solidification, surface pressure, or vertical stratification of the atmosphere, to be intimately controlled by the dominant volatile and outgassing history of the planet. Thermal sequences fall into three general classes with increasing cooling timescale: CO, N$_2$, and O$_2$ with minimal effect, H$_2$O, CO$_2$, and CH$_4$ with intermediate influence, and H$_2$ with several orders of magnitude increase in solidification time and atmosphere vertical stratification. Our numerical experiments exemplify the capabilities of the presented modeling framework and link the interior and atmospheric evolution of rocky exoplanets with multi-wavelength astronomical observations.

astro-ph.EP↗

arXiv2019

Linking the evolution of terrestrial interiors and an early outgassed atmosphere to astrophysical observations

A terrestrial planet is molten during formation and may remain so if subject to intense insolation or tidal forces. Observations continue to favour the detection and characterisation of hot planets, potentially with large outgassed atmospheres. We aim to determine the radius of hot Earth-like planets with large outgassed atmospheres and explore differences between molten and solid silicate planets and their influence on the mass-radius relationship and transmission and emission spectra. An interior-atmosphere model, combined with static structure calculations, tracks the evolving radius of a rocky mantle that is outgassing CO$_2$ and H$_2$O. Synthetic emission and transmission spectra are generated for CO$_2$ and H$_2$O dominated atmospheres. Atmospheres dominated by CO$_2$ suppress the outgassing of H$_2$O to a greater extent than previously realised, as previous studies have applied an erroneous relationship between volatile mass and partial pressure. We therefore predict more H$_2$O can be retained by the interior during the later stages of magma ocean crystallisation. Furthermore, formation of a lid at the surface can tie outgassing of H$_2$O to the efficiency of heat transport through the lid, rather than the atmosphere's radiative timescale. Contraction of the mantle as it solidifies gives $\sim5\%$ radius decrease, which can partly be offset by addition of a relatively light species to the atmosphere. We conclude that a molten silicate mantle can increase the radius of a terrestrial planet by around $5\%$ compared to its solid counterpart, or equivalently account for a $13\%$ decrease in bulk density. An outgassing atmosphere can perturb the total radius according to its speciation. Atmospheres of terrestrial planets around M-stars that are dominated by CO$_2$ or H$_2$O can be distinguished by observing facilities with extended wavelength coverage (e.g., JWST).

astro-ph.EP↗

arXiv2017

Numerical solution of a non-linear conservation law applicable to the interior dynamics of partially molten planets

The energy balance of a partially molten rocky planet can be expressed as a non-linear diffusion equation using mixing length theory to quantify heat transport by both convection and mixing of the melt and solid phases. In this formulation the effective or eddy diffusivity depends on the entropy gradient, $\partial S/\partial r$, as well as entropy. First we present a simplified model with semi-analytical solutions, highlighting the large dynamic range of $\partial S/\partial r$, around 12 orders of magnitude, for physically-relevant parameters. It also elucidates the thermal structure of a magma ocean during the earliest stage of crystal formation. This motivates the development of a simple, stable numerical scheme able to capture the large dynamic range of $\partial S/\partial r$ and provide a flexible and robust method for time-integrating the energy equation. We then consider a full model including energy fluxes associated with convection, mixing, gravitational separation, and conduction that all depend on the thermophysical properties of the melt and solid phases. This model is discretised and evolved by applying the finite volume method (FVM), allowing for extended precision calculations and using $\partial S/\partial r$ as the solution variable. The FVM is well-suited to this problem since it is naturally energy conserving, flexible, and intuitive to incorporate arbitrary non-linear fluxes that rely on lookup data. Special attention is given to the numerically challenging scenario in which crystals first form in the centre of a magma ocean. Our computational framework is immediately applicable to modelling high melt fraction phenomena in Earth and planetary science research. Furthermore, it provides a template for solving similar non-linear diffusion equations arising in other disciplines, particularly for non-linear functional forms of the diffusion coefficient.

astro-ph.EP↗

arXiv2016

Extreme-scale Multigrid Components within PETSc

Elliptic partial differential equations (PDEs) frequently arise in continuum descriptions of physical processes relevant to science and engineering. Multilevel preconditioners represent a family of scalable techniques for solving discrete PDEs of this type and thus are the method of choice for high-resolution simulations. The scalability and time-to-solution of massively parallel multilevel preconditioners can be adversely effected by using a coarse-level solver with sub-optimal algorithmic complexity. To maintain scalability, agglomeration techniques applied to the coarse level have been shown to be necessary. In this work, we present a new software component introduced within the Portable Extensible Toolkit for Scientific computation (PETSc) which permits agglomeration. We provide an overview of the design and implementation of this functionality, together with several use cases highlighting the benefits of agglomeration. Lastly, we demonstrate via numerical experiments employing geometric multigrid with structured meshes, the flexibility and performance gains possible using our MPI-rank agglomeration implementation.

cs.MS↗

arXiv2016

A Stochastic Performance Model for Pipelined Krylov Methods

Pipelined Krylov methods seek to ameliorate the latency due to inner products necessary for projection by overlapping it with the computation associated with sparse matrix-vector multiplication. We clarify a folk theorem that this can only result in a speedup of $2\times$ over the naive implementation. Examining many repeated runs, we show that stochastic noise also contributes to the latency, and we model this using an analytical probability distribution. Our analysis shows that speedups greater than $2\times$ are possible with these algorithms.

cs.DC↗

arXiv2015

Pipelined, Flexible Krylov Subspace Methods

We present variants of the Conjugate Gradient (CG), Conjugate Residual (CR), and Generalized Minimal Residual (GMRES) methods which are both pipelined and flexible. These allow computation of inner products and norms to be overlapped with operator and nonlinear or nondeterministic preconditioner application.The methods are hence aimed at hiding network latencies and synchronizations which can become computational bottlenecks in Krylov methods on extreme-scale systems or in the strong-scaling limit. The new variants are not arithmetically equivalent to their base flexible Krylov methods, but are chosen to be similarly performant in a realistic use case, the application of strong nonlinear preconditioners to large problems which require many Krylov iterations. We provide scalable implementations of our methods as contributions to the PETSc package and demonstrate their effectiveness with practical examples derived from models of mantle convection and lithospheric dynamics with heterogeneous viscosity structure. These represent challenging problems where multiscale nonlinear preconditioners are required for the current state-of-the-art algorithms, and are hence amenable to acceleration with our new techniques. Large-scale tests are performed in the strong-scaling regime on a contemporary leadership supercomputer, where speedups approaching, and even exceeding $2\times$ can be observed. We conclude by analyzing our new methods with a performance model targeted at future exascale machines.

math.NA↗