In this section, we will explain how to launch a CalculiX job. You can choose to run the job on your system, either Linux or Windows, or use a remote cloud HPC.

Hereafter, each of these methods will be explained:

• Download a pre-compiled version for Linux or Windows OS and set up the correct environment in order to launch the simulation.

  This link is for the Linux executable: CalculiX

  This link is for the Windows executable: CalculiX
  A complete tutorial on how to correctly set up the Windows environment will be published soon …

• Use a cloud HPC. We advice you two solutions: the first one FEA Cluster, and the second one CLOUD HPC provided by CFD FEA/FEM service.

Using FEA Cluster, you can submit your job for free but with some restrictions. Simply upload your model in any common compressed format (.gz, .bz2, .Z, .zip, .zipx, .tar, .7z) with a maximum file size limit of 150MB and submit the job. After the job is completed, you will receive an email notification with the possibility to download the results within a time limit of 24 hours. Another limitation is the execution time, which varies based on the number of selected CPUs and memory usage.

They offer upto 30 cpus and 256 GB RAM per machine. They feature Intel Scalable Processors (Cascade Lake) and up to 3.8Ghz sustained all-core-turbo.

Using the CLOUD HPC you can submit your job without any restrictions on the number of CPUs or memory usage. However, this service is subject to a pricing policy, which is explained on this page: Pricing.

Cloud computing is the way through which CFD FEA SERVICE want to deliver complex hardware systems dedicated to engineering simulations. CFD FEA SERVICE have created an innovative cloud High Performance Computing (HPC) system for the Computer Aided Engineering (CAE) world.

Thanks to this cloud HPC you can use the power of a cluster to run your engineering analyses, may them be CFD, FEM, FEA and many more. The system allows you to select machines in terms of available vCPU and RAM to solve the case under study.

If you are you looking for an answer to your fluctuating hardware resource requirements, if you would you like to enjoy pay-per-use pricing and you need to scale your computational capacity then CFD FEA SERVICE cloud HPC service has you covered: use our system to enjoy 360° services from cloud resources and clusters to web accessibility with a wide range of integrated and always up to date softwares.

Moreover Cloud computing offer a professional SERVICE LEVEL AGREEMENT to its costumer.

The memory usage in implicit FEM simulation is very high. I still remember the 2GB limit of 32-bit architecture machines. Nowadays, the latest 64-bit hardware in laptops or desktop PCs can easily reach 64GB or more of installed RAM. Despite this large amount of memory being considered substantial, the memory consumption grows exponentially with the size of your FEM model in terms of Degrees of Freedom, which is directly correlated to the number of nodes.

Have a look at the graph below to see an indication of how the memory consumption is correlated to the size of your FEM model:

Nowadays, most computers have a single socket with several cores, allowing calculations to be performed in parallel. To perform CalculiX calculations in parallel, you need to set the environment variable OMP_NUM_THREADS or the environment variable CCX_NPROC_STIFFNESS to the desired number of cores you want to use.

The table below shows the CPU time versus the number of cores and for different CalculiX solver used (SPOOLES and PARDISO):