![]() ![]() ![]() For the transfer of volumes, the user can check the Crossline and/or Time-Slice options to associate the X-line and Time-slide volumes in the copy. If the project is not listed, they can search for it using the Scan Projects button.Data can be transferred between the current project and a remote project. To display the remote project, the user has to select it from the pull-down menu. In the Project Data Exchange window, the current project is displayed on the left-hand side of the window, whereas the remote project is on the right-hand side. Note that it is impossible to delete the Inline volume via this interface because all the other volumes (X-Line and Time Slice) are computed from the Inline one. Finally, they can click on Run to launch the computations. The user has to check or uncheck the boxes to either compute or delete the associated volume. The already computed volumes are in green. The volumes that have not been computed are in red. The total size of the project is displayed at the bottom left corner of the window. The Data Management tool is a table which quickly summarizes the X-line and Time Slice volumes and their sizes. To reduce the size of a project, the user can use the Data management tool available from the general toolbar. The Model-Grid can then be created on this cropped volume. Using the Extraction tools available from the Volume toolbar, the user can create an extraction area and then extract this sub volume from the original one. This preliminary data conditioning can be done using the Vertical Shift tool available from the Horizon toolbar.Ī cropped seismic volume can also be created. In order to properly interpret and model the stratigraphic level of interest, it is strongly advised to shift both the upper and lower boundaries of +/- 100 to 300ms. The Propagate/Interpolate tool available from the Horizon toolbar can be used to fill existing gaps or to generate full surfaces from a picked horizon. The input horizons need to be continuous where the Model-Grid is to be generated: no node will be computed between a hole (=void) and a surface, or a hole (=void) and another hole (=void). The option is available in the Advanced Options of the Model-Grid Creation tool. The grid can be generated between two horizons, or from the top/bottom of the seismic to the key horizon. Our internal benchmarks show that PaleoScan’s performance does not increase linearly with the number of cores used: 8 cores will definitely improve the performance but it will peak at 16 cores and even decrease beyond 16. Distributed environments, with several physical machines with separated memory and CPU, are therefore not supported. Paleoscan’s tasks (Geo-Model, attributes computation.) can only be started from the PaleoScan GUI (Graphical User Interface) as single executables. Threads cannot be processed separately on different CPUs and memories. However, all the threads are processed on the machine running PaleoScan and share the machine’s memory. The other processes that do not support GPU but CPU parallelization run faster when CPU multithreading (several threads running on the same CPU) is enabled. When both are enabled, GPU is used instead of CPU for the processes that support GPU. Note that enabling the GPU parallelization will not dramatically increase Paleoscan’s performance except for the Properties Modeling module.ĬPU parallelization is widely used in PaleoScan but GPU acceleration (using OpenCL technology) is currently used only in the Properties Modeling module and for some attributes computation, such as similarity. The recommendation for GPU usage is to select the graphic card and for the CPU usage is to keep the default number of threads. ![]()
0 Comments
Leave a Reply. |