Search for: All records

We present a novel machine learning (ML)-based method to accelerate conservative-to-primitive inversion, focusing on hybrid piecewise polytropic and tabulated equations of state. Traditional root-finding techniques are computationally expensive, particularly for large-scale relativistic hydrodynamics simulations. To address this, we employ feedforward neural networks (NNC2PS and NNC2PL), trained in PyTorch (2.0+) and optimized for GPU inference using NVIDIA TensorRT (8.4.1), achieving significant speedups with minimal accuracy loss. The NNC2PS model achieves L1 and L∞ errors of 4.54×10−7 and 3.44×10−6, respectively, while the NNC2PL model exhibits even lower error values. TensorRT optimization with mixed-precision deployment substantially accelerates performance compared to traditional root-finding methods. Specifically, the mixed-precision TensorRT engine for NNC2PS achieves inference speeds approximately 400 times faster than a traditional single-threaded CPU implementation for a dataset size of 1,000,000 points. Ideal parallelization across an entire compute node in the Delta supercomputer (dual AMD 64-core 2.45 GHz Milan processors and 8 NVIDIA A100 GPUs with 40 GB HBM2 RAM and NVLink) predicts a 25-fold speedup for TensorRT over an optimally parallelized numerical method when processing 8 million data points. Moreover, the ML method exhibits sub-linear scaling with increasing dataset sizes. We release the scientific software developed, enabling further validation and extension of our findings. By exploiting the underlying symmetries within the equation of state, these findings highlight the potential of ML, combined with GPU optimization and model quantization, to accelerate conservative-to-primitive inversion in relativistic hydrodynamics simulations. 

We present a novel machine learning (ML) method to accelerate conservative-to-primitive inversion, focusing on hybrid piecewise polytropic and tabulated equations of state. Traditional root-finding techniques are computationally expensive, particularly for large-scale relativistic hydrodynamics simulations. To address this, we employ feedforward neural networks (NNC2PS and NNC2PL), trained in PyTorch and optimized for GPU inference using NVIDIA TensorRT, achieving significant speedups with minimal accuracy loss. The NNC2PS model achieves L1 and L∞ errors of 4.54×10−7 and 3.44×10−6, respectively, while the NNC2PL model exhibits even lower error values. TensorRT optimization with mixed-precision deployment substantially accelerates performance compared to traditional root-finding methods. Specifically, the mixed-precision TensorRT engine for NNC2PS achieves inference speeds approximately 400 times faster than a traditional single-threaded CPU implementation for a dataset size of 1,000,000 points. Ideal parallelization across an entire compute node in the Delta supercomputer (Dual AMD 64 core 2.45 GHz Milan processors; and 8 NVIDIA A100 GPUs with 40 GB HBM2 RAM and NVLink) predicts a 25-fold speedup for TensorRT over an optimally-parallelized numerical method when processing 8 million data points. Moreover, the ML method exhibits sub-linear scaling with increasing dataset sizes. We release the scientific software developed, enabling further validation and extension of our findings. This work underscores the potential of ML, combined with GPU optimization and model quantization, to accelerate conservative-to-primitive inversion in relativistic hydrodynamics simulations. 

The prediction of protein 3D structure from amino acid sequence is a computational grand challenge in biophysics and plays a key role in robust protein structure prediction algorithms, from drug discovery to genome interpretation. The advent of AI models, such as AlphaFold, is revolutionizing applications that depend on robust protein structure prediction algorithms. To maximize the impact, and ease the usability, of these AI tools we introduce APACE, AlphaFold2 and advanced computing as a service, a computational framework that effectively handles this AI model and its TB-size database to conduct accelerated protein structure prediction analyses in modern supercomputing environments. We deployed APACE in the Delta and Polaris supercomputers and quantified its performance for accurate protein structure predictions using four exemplar proteins: 6AWO, 6OAN, 7MEZ, and 6D6U. Using up to 300 ensembles, distributed across 200 NVIDIA A100 GPUs, we found that APACE is up to two orders of magnitude faster than off-the-self AlphaFold2 implementations, reducing time-to-solution from weeks to minutes. This computational approach may be readily linked with robotics laboratories to automate and accelerate scientific discovery. 

Abstract We presentAsterX, a novel open-source, modular, GPU-accelerated, fully general relativistic magnetohydrodynamic (GRMHD) code designed for dynamic spacetimes in 3D Cartesian coordinates, and tailored for exascale computing. We utilize block-structured adaptive mesh refinement (AMR) throughCarpetX, the new driver for theEinstein Toolkit, which is built onAMReX, a software framework for massively parallel applications.AsterXemploys the Valencia formulation for GRMHD, coupled with the ‘Z4c’ formalism for spacetime evolution, while incorporating high resolution shock capturing schemes to accurately handle the hydrodynamics.AsterXhas undergone rigorous testing in both static and dynamic spacetime, demonstrating remarkable accuracy and agreement with other codes in literature. Using subcycling in time, we find an overall performance gain of factor 2.5–4.5. Benchmarking the code through scaling tests on OLCF’s Frontier supercomputer, we demonstrate a weak scaling efficiency of about 67%–77% on 4096 nodes compared to an 8-node performance. 

Triaxial neutron stars can be sources of continuous gravitational radiation detectable by ground-based interferometers. The amplitude of the emitted gravitational wave can be greatly affected by the state of the hydrodynamical fluid flow inside the neutron star. In this work, we examine the most triaxial models along two sequences of constant rest mass, confirming their dynamical stability. We also study the response of a triaxial figure of quasiequilibrium under a variety of perturbations that lead to different fluid flows. Starting from the general relativistic compressible analog of the Newtonian Jacobi ellipsoid, we perform simulations of Dedekind-type flows. We find that in some cases the triaxial neutron star resembles a Riemann-S-type ellipsoid with minor rotation and gravitational wave emission as it evolves towards axisymmetry. The present results highlight the importance of understanding the fluid flow in the interior of a neutron star in terms of its gravitational wave content. 

The Einstein Toolkit is a complex software system for numerical general relativity, a science domain that includes colliding black holes, neutron stars, supernovae, etc. As might be expected for a framework of this size and age (parts of it are over 20 years old), there is a significant learning curve to building it, running it, writing new modules for it, etc. Over the years, the Einstein Toolkit maintainers have given a number of tutorials for new users. In recent years, we have created a tutorial server which allows us to streamline the teaching/learning process through the use of Jupyter notebooks and docker images. In this paper we describe the special considerations and adaptations required by the image and the notebook server that enable us to (1) easily make logins and manage accounts which streamlines both the classroom and the independent study experiences, (2) create a simplified but natural user experience for compiling and developing a complex C++ application, (3) scale to increasing class sizes. 

Reproducibility of results is a cornerstone of the scientific method. Scientific computing encounters two challenges when aiming for this goal. Firstly, reproducibility should not depend on details of the runtime environment, such as the compiler version or computing environment, so results are verifiable by third-parties. Secondly, different versions of software code executed in the same runtime environment should produce consistent numerical results for physical quantities. In this manuscript, we test the feasibility of reproducing scientific results obtained using the IllinoisGRMHD code that is part of an open-source community software for simulation in relativistic astrophysics, the Einstein Toolkit. We verify that numerical results of simulating a single isolated neutron star with IllinoisGRMHD can be reproduced, and compare them to results reported by the code authors in 2015. We use two different supercomputers: Expanse at SDSC, and Stampede2 at TACC. By compiling the source code archived along with the paper on both Expanse and Stampede2, we find that IllinoisGRMHD reproduces results published in its announcement paper up to errors comparable to round-off level changes in initial data parameters. We also verify that a current version of IllinoisGRMHD reproduces these results once we account for bug fixes which have occurred since the original publication.