EAGE Seventh High Performance Computing Workshop
- Conference date: September 25 - 27, 2023
- Location: Lugano, Switzerland
- Published: 25 September 2023
-
-
A Closer Look at the Opportunities for Analogue Quantum Computing in Future Upstream HPC Applications
More LessAuthors M. Dukalski and M. MöllerSummaryIn this work we closely examine the overlap between (1) the computationally challenging problems in geophysical applications in the upstream business and (2) the capabilities of quantum annealing (QA) quantum processing units (QPUs). We analyze the strengths and limitations of the latter and we explain how and why QPUs need to be assisted with CPUs and GPUs to solve combinatorial optimization problems at industrial scale. We illustrate the discussion with three potential use cases: stack power maximization for residual statics estimation, “ray tracing” and compressive sensing.
-
-
-
A Reservoir Model Characterization with a Bayesian Framework and a Modulus based Physics constrained Neural Operator
More LessAuthors C. Etienam, I. Said, O. Ovcharenko and K. HesterSummaryWe have developed a novel inverse methodology workflow using Nvidia Modulus’s physics-informed neural operator (PINO) as a fast surrogate to solve the black-oil forward problem together with a modified Bayesian adaptive ensemble Kalman inversion (aREKI) approach parametrized with deep neural network (DNN) exotic priors. The DNN prior used here are the variational convolution autoencoder and a novel mixture of deep neural network experts, and together with the developed aREKI method is optimal for sampling non-Gaussian posterior measures, like situations found in channelized reservoirs. The output from the PINO model is the pressure and water saturation fields, and the inputs are the absolute permeability, effective porosity,time-step,initial pressure & saturation, and source and sink terms. The overall workflow is successful in recovering the unknown channelized absolute permeability & effective porosity field for this synthetic case, using an ensemble of 5,000 members with about 100X speedup to traditional numerical approaches. The workflow is suitable for forward and inverse reservoir uncertainty quantification tasks for rapid reservoir management operational decision strategies.
-
-
-
Improved GPU-based Full Waveform Inversion Algorithm with a Lossy Compression and Multiple CUDA Streams
More LessAuthors Y.S. Kim, M. Dmitriev and H. SalimSummaryFull Waveform Inversion (FWI) is a wave-based solution that can establish an accurate velocity model. Such a model is crucial for generating high-fidelity seismic images that aid in the discovery of reservoirs. However, FWI has the disadvantage that it requires considerable computing costs. These computational costs can be dramatically high in case of large-scale models with fine grid sizes. To resolve this issue, we optimize an original FWI workflow more adequate to a graphic processing unit (GPU) system. First, most of FWI functions are written with the compute unified device architecture (CUDA) program language. Second, we employ a GPU-based lossy compression library to reduce the memory requirements for storing source wavefields. It can help us to save wavefields into computer memory instead of hard disk. Then, we modified an equation for the zero-lag cross correlation for the gradient vector calculation in order to utilize multiple CUDA streams. To verify our optimized FWI workflow, we compare the performance speed between the original and our proposed FWI algorithms in the various aspects.
-
-
-
Chronos: a GPU-accelerated and Energy Efficient Linear Solver for Large Scale Simulations in Geoscience
More LessAuthors C. Janna, M. Frigo, G. Isotton, N. Spiezia and D. ColomboSummaryNumerical simulations are widely used in studying underground processes, as direct measurements are extremely expensive or even unfeasible. Typical examples are the geomechanical processes involved in compacting reservoirs, the evolution of sedimentary basins or the fluid flows in deep, porous formations. Developing efficient, robust and scalable linear solvers to solve those problems is a crucial task. Graphics Processing Units (GPUs) are attracting a growing attention since they are well suited for massively parallel computations providing a very good balance among performance, price and power consumption. In the field of iterative methods, for instance, it is difficult to take advantage from preconditioners based on incomplete factorization and approximate inverses are generally preferred.
In this work, we will focus on the adaptive Factored Sparse Approximate Inverses (aFSAI) that have been already successfully tested on GPUs showing significant speed-ups with respect to the CPU counterpart. We will show in problems arising from geomechanics and basin modelling that, thanks to this GPU- accelerated, it is possible not only to speed-up the simulations but also to significantly reduce the energy consumption to power the hardware.
-
-
-
High-dimension Seismic Data Regularization on CPU/GPU Heterogeneous Computing System
More LessSummarySeismic data is usually acquired irregularly and sparsely due to obstacles and high cost of acquisition, making the regularization of seismic data with anti-leakage and anti-aliasing techniques essential in the seismic data processing workflow. We propose a new method of 4D anti-leakage and anti-aliasing Fourier reconstruction using a cube-removal strategy to handle the combination of irregular sampling and aliasing in high-dimension seismic data. Our approach computes a weighting function by stacking the spectrum along the radial lines to suppress the aliasing energy, and then iteratively pick the largest amplitude cube to construct the Fourier spectrum. To fully leverage the power of the supercomputer, we design a multiple-level parallel architecture using CPU/GPU heterogeneous computing system. Specifically, we have developed parallel data splitting, multiple GPU devices computing, and parallel merging. The efficiency test validates that the parallel architecture can achieve a high acceleration ratio compared with CPU version and attain nearly linear performance scalability with the GPU devices number. The numerical test on one field data example demonstrates the robustness and effectiveness of our method.
-
-
-
Seismic Data Compression in Time Processing
More LessAuthors M. Dmitriev, T. Tonellot, Y.S. Kim, M. Almubarak and H.J. AlSalemSummaryModern seismic data acquisition with high-channel count or point-source and point-receiver recording systems can produce massive volumes of data, often exceeding hundreds of terabytes for a single survey. Processing such vast volumes of data becomes challenging due to the limited I/O bandwidth and storage capacity. The use of seismic data compression can help to address these challenges and facilitate more efficient data management. In this paper, we evaluate several cutting-edge lossy compressors, including SZ3, ZFP, and Bitcomp, for compressing seismic data. We compare compression ratios achieved by all compressors using different user-defined error bounds, and explore how compression errors can affect the accuracy of the reconstructed pre-stack gathers and final stacked image. We also examine the difference in compression efficiency between 1D and 2D techniques.
-
-
-
A Scalable Cloud-Native Computational Framework with Applications in Wind Farm Optimization
More LessAuthors V. Ananthan, S. Tadepalli and A. RasheedSummaryThis work introduces a cloud-native computational framework, called TwinGraph, for large scale high-throughput and high-performance computing. The capability of TwinGraph to orchestrate complex, algorithmically defined workflows for engineering design optimization through simulation models which can range in the underlying fidelity is shown; in so doing, TwinGraph is demonstrated to address technical challenges in the energy industry related to high-performance simulation workflows at scale in cloud and hybrid environments. The challenges in designing such frameworks, as well as the underlying architecture of the framework are also presented. Subsequently, the application of TwinGraph to an example application of wind farm problem is presented, where a stochastically perturbed gradient-descent optimization routine is used on top of a simulation model to find optimal layout configurations for a farm of fixed area. Finally, the future of such frameworks for running large-scale, higher fidelity blade-resolved wind simulations and simulation-driven optimization are discussed.
-
-
-
Massively Distributed Reverse Time Migration
More LessAuthors M. Araya-Polo, M. Jacquelin and D. KlahrSummaryReverse Time Migration (RTM) is the seismic imaging everyday tool. Traditionally, it is implemented with finite differences method, which is at its core a stencil computation. Stencil computations pose major challenges for all major processor architectures with their deep memory hierarchies. In this work, a novel RTM algorithm is presented. It efficiently exploits extreme scale distributed-memory platforms with no cache hierarchy, a paradigm closer to dataflow rather than traditional von Neumann architecture. The implementation following this paradigm delivers disruptive level performance.
-
-
-
Enabling Hydrocarbon Detection based on Large Scale AI
More LessAuthors Y. Nakhate, A. Singh and A. KudarovaSummaryHydrocarbon detection using seismic data involves computationally intensive stages and domain expertise. The large and complex nature of seismic data presents challenges in data transfer, memory usage, and processing time, particularly with pre-stack data. This study introduces an optimized I/O approach to leverage Shell’s Delve deep-learning algorithm for rapid hydrocarbon characterization. Overcoming previous barriers, this approach enables efficient resource exploration and decision-making in the energy sector. The method details the methodology, theory, and conclusions, highlighting the potential of deep learning in analyzing pre-stack seismic data.
-
-
-
From High Frequency RTM to AI/ML Aided Salt Interpretation-harness the Power of the Cloud
More LessAuthors K. Jiao, W. Han, J. Chenin, C. Lagrange, M. Isernia and S. TadepalliSummarySeismic imaging and interpretation are important workflows for the Oil and Gas (O&G) industry, offering critical subsurface insights for informed decision-making. This paper presents a case study of SEAM Subsalt TTI Model that effectively harnesses the advanced capabilities of HPC, cloud platforms, and AI/ML aided interpretation to improve and complement the traditional O&G industry workflow processes.
-
-
-
Quantum Computing with Neutral Atoms with Applications in Energy
More LessAuthors H. SadeghiSummaryThis talk explores the potential of quantum computing and the current market status. It focuses on PASQAL’s technology using neutral atoms for scalable quantum computing. The main categories of algorithms designed for neutral atom quantum computers and their real-world applications are summarized.
The energy industry has numerous quantum applications, including energy resource management, grid optimization, complex system optimization, simulation, and battery design. Three disruptive examples are highlighted.
Integration of quantum hardware with HPC centers is crucial. The talk addresses challenges, the expected maturity of quantum computing, and the integration challenges with HPC centers. It prompts us to consider the future of HPC centers in the era of quantum computing.
-
-
-
Generative Modeling and Inverse Mapping for Reservoir Navigation with Shallow-to-Ultra-Deep LWD Electromagnetics
More LessAuthors A. Cheryauka, D. Safin, A. Vianna, E. Ferreira and W. FernandesSummaryWe have tested our pilot 3D generative modeling and hybrid physics + AI/ML mapping applied to oil & gas reservoir navigation. Initially, we developed the capability to statistically mimic a logging survey and analyze the computed electromagnetic tool responses. Then, under real-time and pre-/post-well memory workflow conditions, probabilistic inverse mapping has been investigated aiming at optimal placement of the new wells. We wrap-up our forward and inverse applications into deployable soft notebooks and power them with CPU/GPU parallel computing and 3D scene visualization. The use of supercomputing-on-chip performance features, web-based GUIs and OS-agnostic environments makes these soft notebooks very suitable for fast-fail-fast-learn interactions not only with internal subject-matter experts, but also with early adopters among the partners and customers in the industry.
-
-
-
HBM Contribution to Intel Sapphire Rapids for Geophysical Workloads
More LessAuthors V. ArslanSummaryIn this presentation we are going to show the benefits of new Intel Sapphire Rapids CPU with High-Bandwidth Memory (HBM) applied to state-of-the-art Geophysical application.
-
-
-
HPC Implementation of the FWI Gradient Reduction Stage
More LessAuthors L. Bortot, N. Bienati and J. PanizzardiSummaryDespite its simplicity, the gradient reduction stage can become a bottleneck in the execution of FWI (and RTM as well). Leveraging the file system for this stage can lead to very simple and easy to develop architectures, but performance can be limited and variability of performance can be significant. Leveraging the network may increase perfromance and reduce varibility, but the complexity to be managed is relevant. In this abstract, the optimization of this stage is analyzed and discussed
-
-
-
Benchmarking Simulated and Physical Quantum Processing Units Using Quantum and Hybrid Algorithms
More LessAuthors A. MelnikovSummaryQuantum computing is a rapidly growing technology field with increasingly useful applications across industry and research. This new paradigm of computing has the potential to solve classically intractable problems by exploiting an exponentially increasing computational space. This allows quantum algorithms to dramatically reduce the runtime for solving computationally resource-intensive problems. Quantum Neural Networks (QNNs) [ 1 - 4 ] present a promising opportunity for overcoming the scaling problem in classical machine learning. However, training QNNs requires significant circuit evaluations, making them a resource-intensive use case for quantum processing units (QPUs). This study compares software development kits (SDKs) and hardware platforms to determine the fastest and most cost-efficient combination for developing novel quantum algorithms and uses QNNs as a general benchmark. The aim is to find the optimal balance of runtime, cost, and accuracy for executing quantum circuits.
-
-
-
Visco-Elastic Wave Propagation on GPUs
More LessAuthors S. Reker, A. St-Cyr, D. Cha, S. Geevers, C. Vosenek, M. Bosmans, T. Vuik, D. Van Eijkeren, M. Van der Kolk, J. Van der Holst, S. Banerjee and M. Van der VeenSummaryWith the advent of large offset and low frequency seismic data, the information stored in surveys has become ever richer and more voluminous. At the same time, a push for more detailed solutions requires the inclusion of higher frequencies from the data. Moreover, to support extracting accurate and realistic geophysical models of the subsurface, velocity model building such as done in Full Waveform Inversion (FWI) frequently requires inclusion of anisotropic parameters, elastic, and viscous information. However, the computational cost associated with solving such realistic equations is non-trivial.
In the current work, we have implemented the fully anisotropic viscoelastic equations (and all simpler cases). We use a velocity-stress staggered grid approach proposed in [ 5 ] with optimized FD weights of any spatial order and second order in time. We discuss performance for multi-GPUs & multi-node GPUs, as well as some of the optimization techniques used for convolutional perfectly matched layer (CPML) and MPI.
-
-
-
Cross-platform Benchmarking of Seismic Imaging Kernels
More LessAuthors G. Gorman, L. Decker, A. Loddoch, M. Louboutin, F. Luporini, R. Nelson, M. Roberts, A. St-Cyr and J. TillaySummaryThe article presents a framework designed for standardized cross-platform benchmarking of seismic imaging kernels. This framework aims to standardise the benchmarking process, providing reference implementations for various architectures and enabling reproducible experiments. The system uses GitHub Actions for automation, a development cluster composed of different computer architectures, and on-demand runners on AWS. This setup allows for easy comparison of different methods, hardware, while also enabling the integration of more tools and servers. Preliminary results are presented for 3D isotropic acoustic and Fletcher Du Fowler TTI benchmarks, using Giga-points per second (GPts/s) as a performance unit, providing more practical insights compared to the standard FLOPS (Floating Point Operations Per Second) metric.
-
-
-
Efficient Execution of MPI Containers
More LessAuthors P. Souza Filho, A. Bulcão, J. Panetta, M. Lough and B. MonnetSummaryWe propose a runtime method to partial or fully override the MPI inside the container, by one version that is optimized for the target machine. Our approach does not require a container image rebuild/update and doesn’t require a match between the host and the container OS. We executed a high order 3D stencil using two nodes with two MPI processes per node (PPN) to demonstrate the performance difference by the original container with Intel MPI, and an overridden container with Cray and MVAPICH2 tuned for the target machine Slingshot fabric.
-