Guofeng Liu

3D seismic reverse time migration on GPGPU

Reverse time migration (RTM) is a powerful seismic imaging method for the interpretation of steep-dips and subsalt regions; however, implementation of the RTM method is computationally expensive. In this paper, we present a fast and computationally inexpensive implementation of RTM using a NVIDIA general purpose graphic processing unit (GPGPU) powered with Compute Unified Device Architecture (CUDA). To accomplish this, we introduced a random velocity boundary in the source propagation kernel. By creating a random velocity layer at the left, right, and bottom boundaries, the wave fields that encounter the boundary regions are pseudo-randomized. Reflections off the random layers have minimal coherent correlation in the reverse direction. This process eliminates the need to write the wave fields to a disk, which is important when using a GPU because of the limited bandwidth of the PCI-E that is connected to the CPU and GPU. There are four GPU kernels in the code: shot, receiver, modeling, and imaging. The shot and receiver insertion kernels are simple and are computed using a GPU because the wave fields reside in GPUs memory. The modeling kernel is computed using Micikeviciuss tiling method, which uses shared memory to improve bandwidth usage in 2D and 3D finite difference problems. In the imaging kernel, we also use this tiling method. A Tesla C2050 GPU with 4GB memory and 480 stream processing units was used to test the code. The shot and receiver modeling kernel occupancy achieved 85%, and the imaging kernel occupancy was 100%. This means that the code achieved a good level of optimization. A salt model test verified the correct and effective implementation of the GPU RTM code.

GICUDA: A parallel program for 3D correlation imaging of large scale gravity and gravity gradiometry data on graphics processing units with CUDA

The 3D correlation imaging for gravity and gravity gradiometry data provides a rapid approach to the equivalent estimation of objective bodies with different density contrasts in the subsurface. The subsurface is divided into a 3D regular grid, and then a cross correlation between the observed data and the theoretical gravity anomaly due to a point mass source is calculated at each grid node. The resultant correlation coefficients are adopted to describe the equivalent mass distribution in a quantitate probability sense. However, when the size of the survey data is large, it is still computationally expensive. With the advent of the CUDA, GPUs lead to a new path for parallel computing, which have been widely applied in seismic processing, astronomy, molecular dynamics simulation, fluid mechanics and some other fields. We transfer the main time-consuming program of 3D correlation imaging into GPU device, where the program can be executed in a parallel way. The synthetic and real tests have been performed to validate the correctness of our code on NVIDIA GTX 550. The precision evaluation and performance speedup comparison of the CPU and GPU implementations are illustrated with different sizes of gravity data. When the size of grid nodes and observed data sets is 1024x1024x1 and 1024x1024, the speed up can reach to 81.5 for gravity data and 90.7 for gravity vertical gradient data respectively, thus providing the basis for the rapid interpretation of gravity and gravity gradiometry data.

A practical implementation of 3D TTI reverse time migration with multi-GPUs

Tilted transversely isotropic (TTI) media are typical earth anisotropy media from practical observational studies. Accurate anisotropic imaging is recognized as a breakthrough in areas with complex anisotropic structures. TTI reverse time migration (RTM) is an important method for these areas. However, P and SV waves are coupled together in the pseudo-acoustic wave equation. The SV wave is regarded as an artifact for RTM of the P wave. We adopt matching of the anisotropy parameters to suppress the SV artifacts. Another problem in the implementation of TTI RTM is instability of the numerical solution for a variably oriented axis of symmetry. We adopt Fletchers equation by setting a small amount of SV velocity without an acoustic approximation to stabilize the wavefield propagation. To improve calculation efficiency, we use NVIDIA graphic processing unit (GPU) with compute unified device architecture instead of traditional CPU architecture. To accomplish this, we introduced a random velocity boundary and an extended homogeneous anisotropic boundary for the remaining four anisotropic parameters in the source propagation. This process avoids large storage memory and IO requirements, which is important when using a GPU with limited bandwidth of PCI-E. Furthermore, we extend the single GPU code to multi-GPUs and present a corresponding high concurrent strategy with multiple asynchronous streams, which closely achieved an ideal speedup ratio of 2:1 when compared with a single GPU. Synthetic tests validate the correctness and effectiveness of our multi-GPUs-based TTI RTM method. We provide a practical solution for 3D TTI RTM using Multi-GPUs.We introduce a modified random boundary conditions suitable for TTI RTM.We adopt asynchronous streams to maximize the capacity utilization of the GPU.Our algorithm based achieved an ideal speedup ratio of 2:1 compared with a single GPU.

Review: The RASC and CASC programs for ranking, scaling and correlation of biostratigraphic events

RASC is an acronym for RAnking and SCcaling of biostratigraphic events. Code of the RASC computer program was originally published in Computers & Geosciences. During the past 30 years this program has been continuously maintained and updated. Its purpose is to combine biostratigraphic data from land-based sections or exploratory wells drilled in sedimentary basins to construct a regional biozonation that can be used for correlation between sections within a study area. Different methods of quantitative stratigraphy are briefly reviewed in this paper. Ranking is illustrated by application to a simple, artificial dataset. Scaling is explained as a refinement of ranking. Implications of sampling of stratigraphic sections are discussed in detail. Several examples of past successful large-scale RASC applications are given. These include results for well data from the Cenozoic North Sea basin, northwestern Atlantic margin and the Cretaceous seaway between Norway and Greenland. Paleoceanographic interpretations of RASC biozonations supplemented by analysis of variance and correlations between wells are exemplified as well. History of program development is summarized in an Appendix.

Accelerating finite difference wavefield-continuation depth migration by GPU

The most popular hardware used for parallel depth migration is the PC-Cluster but its application is limited due to large space occupation and high power consumption. In this paper, we introduce a new hardware architecture, based on which the finite difference (FD) wavefield-continuation depth migration can be conducted using the Graphics Processing Unit (GPU) as a CPU coprocessor. We demonstrate the program module and three key optimization steps for implementing FD depth migration: memory, thread structure, and instruction optimizations and consider evaluation methods for the amount of optimization. 2D and 3D models are used to test depth migration on the GPU. The tested results show that the depth migration computational efficiency greatly increased using the general-purpose GPU, increasing by at least 25 times compared to the AMD 2.5 GHz CPU.

Seismic Data Prestack Kirchhoff Time Migration with Multi-GPUs

In this paper, we present a scheme of Prestack Kirchhoff Time Migration(PKTM) with multi-GPU. Firstly, we intorduced three main optimization points of GPU code of PKTM, then we test the code with a real field data. After analysis the efficiency curve, we proposed a multi-GPU flowchart of PKTM. it first splits the seismic data to different GPU nodes according to the offset range and collecting and sorting the result to CRP gather , the proposed method can reach the maximum efficency of GPU PKTM code.

The gravity and seismic sequential inversion and its GPU implementation

In this paper, we introduce a gravity and seismic sequential inversion method to invert for density and velocity together. For the gravity inversion, we use an iterative method based on correlation imaging algorithm; for the seismic inversion, we use the full waveform inversion. The link between the density and velocity is an empirical formula called Gardner equation, for large volumes of data, we use the GPU to accelerate the computation.

De-alias seismic data reconstruction investigation

We propose a new De-alias seismic data reconstruction algorithm, which combines minimum-weighted norm interpolation (MWNI) with band extension via modulation (BEM) method effectively. In the algorithm, we use two stages to finish MWNI reconstruction. First, we reconstruct the low-frequency part of the data spectrum using MWNI method. Then based on the reconstructed low-frequency data, the temporary highfrequency data spectrum is reconstructed by BEM, and the temporary spectrum will be used to calculate weight when we finish last high frequency reconstruction by MWNI method. The reconstruction tests on theoretical data and real field data show the proposed method has high efficiency and accuracy, as well as strong De-alias ability.

Event Model For Inclined Incident Multiples And Spectral Factorization

Firstly, the deduction of prediction equation for inclined incident multiples under horizontally layered condition is reviewed briefly. Secondly, based on that equation, event model for generalized Goupillaud layers is put forward which not only describes primary waves but also multiples, even diffraction waves if using prediction equation for laterally variable velocity. Thirdly, to solve the prediction equation, numerical research for both Kolmogorov and Wilson-Burg-Fomel schemes with helical coordinates is carried out. Although the latter is more accurate, it is to some extent also a time-consuming one. In order to find a quick way, Kolmogorov scheme has been improved so as to develop it into a way for prediction spectral factorization where truncation problem has been tackled.

Gridding of Potential Field Data With Inverse Interpolation

Gridding of irregularly distributed potential field data is one of the first and most crucial steps in data analysis. This paper brings forward to realize the gridding of potential field data by using inverse interpolation based on the preconditioning conjugate-gradient iterative algorithm. The gridding method adopts Gaussian weighted coefficients as the interpolation operator, Laplacian operator as the smooth filter, and the inverse of the smooth filter as the preconditioning operator. Experimental results on the synthetic model and the raw aeromagnetic data show that the method is suitable for the features of geophysical potential field, and is a computationally efficient and robust 2-D method with high accuracy and perfect effects.