There is no load balancing so you have to be careful with the resources available on your machine. Here’s a webinar to show you how, Analyze Your Time-Dependent Data 6 Times Faster. You can use PyTecplot and the Python multiprocessing library to launch multiple instances of Tecplot 360 simultaneously to improve the speed of post processing. In Tecplot 360, when you’re working with isosurfaces, slices, or streamtraces, those are fully parallelized and will use all your CPU cores. But you have to be careful because you may run into resource contention if you have multiple instances running at the same time on the same machine. You can also do it on a single machine using multiple instances of Tecplot 360. So, you can get parallel processing done on an HPC. On an HPC, you could submit a batch job where each node on your HPC would load one time step of your data and generate an image. If you have transient data with one file per time step, you can distribute that work among multiple machines. We don’t do distributed memory parallel processing. That means when you are post-processing, all the CPU cores available on a single machine will be used. Tecplot 360 is a shared memory application that uses OpenGL for rendering, which will use your graphics card. Q&A from the Webinar Is it possible to use multiple CPUs or GPUs to speed up post processing?
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |