Leverage the power of high-performance computing in minutes.
Run the 09_Turbine Chrono Example
In this tutorial we will show you how to use the Inductiva API to run an advanced DualSPHysics case that requires significant computing power.
Objective
The goal of this tutorial is to demonstrate how to run the 09_Turbine use case from the examples/chrono directory, available in the DualSPHysics package.
Prerequisites
Download the DualSPHysics package
Download the required files from DualSPHysics_v5.4.2.zip here. You will be working within this directory and
writing the Inductiva Python script there.
Update the simulation script of the 09_Turbine example
Before running the simulation, you will need to adjust the simulation script located
at examples/chrono/09_Turbine/xCaseTurbine_linux64_GPU.sh.
Make the following adjustments:
- Update the
dirbinvariable: Modify thexCaseTurbine_linux64_GPU.shscript to point to the correct binaries directory:export dirbin=/DualSPHysics/bin/linux/ - Remove user input prompt:
To enable automated execution, delete the final line in the script, which waits for user input:
read -n1 -r -p "Press any key to continue..." key
These modifications will prepare the script for seamless automated execution.
Running Your Simulation
Here is the code required to run a DualSPHysics simulation using the Inductiva API:
"""DualSPHysics example."""
import inductiva
# Allocate cloud machine on Google Cloud Platform
cloud_machine = inductiva.resources.MachineGroup(
provider="GCP",
machine_type="g2-standard-32",
spot=True,
data_disk_gb=200)
# Initialize the Simulator
dualsphysics = inductiva.simulators.DualSPHysics( \
version="5.4.1")
# Run simulation
task = dualsphysics.run(
input_dir="/Path/to/09_Turbine",
shell_script="xCaseTurbine_linux64_GPU.sh",
on=cloud_machine)
# Wait for the simulation to finish and download the results
task.wait()
task.download_outputs()
cloud_machine.terminate()
task.print_summary()
This simulation runs in spot mode on a g2-standard-32 machine, featuring 32 virtual CPUs, 1 NVIDIA L4 GPU, and a 200 GB data disk.
Note: Setting
spot=Trueenables the use of spot machines, which are available at substantial discounts. However, your simulation may be interrupted if the cloud provider reclaims the machine.
When the simulation is complete, we terminate the machine, download the results and print a summary of the simulation as shown below.
Task status: Success
Timeline:
Waiting for Input at 08/05, 16:23:13 0.968 s
In Queue at 08/05, 16:23:14 73.641 s
Preparing to Compute at 08/05, 16:24:28 21.629 s
In Progress at 08/05, 16:24:49 64.281 s
└> 64.128 s bash xCaseTurbine_linux64_GPU.sh
Finalizing at 08/05, 16:25:54 16.345 s
Success at 08/05, 16:26:10
Data:
Size of zipped output: 3.52 GB
Size of unzipped output: 4.92 GB
Number of output files: 2033
Total estimated cost (US$): 0.028 US$
Estimated computation cost (US$): 0.018 US$
Task orchestration fee (US$): 0.010 US$
Note: A per-run orchestration fee (0.010 US$) applies to tasks run from 01 Dec 2025, in addition to the computation costs.
Learn more about costs at: https://inductiva.ai/guides/how-it-works/basics/how-much-does-it-cost
As you can see in the "In Progress" line, the part of the timeline that represents the actual execution of the simulation, the core computation time of this simulation was approximately 64.2 seconds.
Download the Results to Your Local Machine
You can donwload the results to your local machine using Inductiva's CLI:
inductiva tasks download u8v7p1v7wfyvvkyc0iq0s632k
Downloading and decompressing data may take a few minutes, depending on your network connection:
Downloading simulation outputs to inductiva_output/u8v7p1v7wfyvvkyc0iq0s632k/output.zip...
100%|█████████████████████████████████████████████████████████████████████████████| 3.52G/3.52G [04:43<00:00, 12.4MB/s]
Uncompressing the outputs to u8v7p1v7wfyvvkyc0iq0s632k...
The results will be stored in the inductiva_output folder, within a subfolder named after the task. Earlier, we set a variable for the internal directory where all outputs would be stored (dirout), which was instantiated as CaseTurbine_out. Let’s check its contents:
ls -las inductiva_output/u8v7p1v7wfyvvkyc0iq0s632k/outputs/CaseTurbine_out
total 36080
0 drwxr-xr-x@ 22 paulobarbosa staff 704 Apr 9 10:55 .
0 drwxr-xr-x@ 5 paulobarbosa staff 160 Apr 9 10:55 ..
9888 -rw-r--r--@ 1 paulobarbosa staff 5058947 Apr 9 10:54 CaseTurbine.bi4
16 -rw-r--r--@ 1 paulobarbosa staff 4523 Apr 9 10:54 CaseTurbine.out
24 -rw-r--r--@ 1 paulobarbosa staff 10830 Apr 9 10:54 CaseTurbine.xml
9432 -rw-r--r--@ 1 paulobarbosa staff 4827935 Apr 9 10:54 CaseTurbine_All.vtk
1280 -rw-r--r--@ 1 paulobarbosa staff 653967 Apr 9 10:54 CaseTurbine_Bound.vtk
8160 -rw-r--r--@ 1 paulobarbosa staff 4174240 Apr 9 10:54 CaseTurbine_Fluid.vtk
752 -rw-r--r--@ 1 paulobarbosa staff 382901 Apr 9 10:54 CaseTurbine_MkCells.vtk
2488 -rw-r--r--@ 1 paulobarbosa staff 1272363 Apr 9 10:54 CaseTurbine__Actual.vtk
8 -rw-r--r--@ 1 paulobarbosa staff 583 Apr 9 10:54 CaseTurbine_dbg-fillbox.vtk
8 -rw-r--r--@ 1 paulobarbosa staff 1947 Apr 9 10:54 CfgChrono_Scheme.vtk
8 -rw-r--r--@ 1 paulobarbosa staff 854 Apr 9 10:54 CfgInit_Domain.vtk
16 -rw-r--r--@ 1 paulobarbosa staff 4155 Apr 9 10:54 CfgInit_MapCells.vtk
8 -rw-r--r--@ 1 paulobarbosa staff 2371 Apr 9 10:54 Floating_Materials.xml
3880 -rw-r--r--@ 1 paulobarbosa staff 1985884 Apr 9 10:54 Rotor.stl
8 -rw-r--r--@ 1 paulobarbosa staff 983 Apr 9 10:54 Run.csv
104 -rw-r--r--@ 1 paulobarbosa staff 51408 Apr 9 10:54 Run.out
0 drwxr-xr-x@ 504 paulobarbosa staff 16128 Apr 9 10:55 boundary
0 drwxr-xr-x@ 508 paulobarbosa staff 16256 Apr 9 10:55 data
0 drwxr-xr-x@ 1006 paulobarbosa staff 32192 Apr 9 10:55 particles
0 drwxr-xr-x@ 503 paulobarbosa staff 16096 Apr 9 10:55 surface
The folders boundary, particles, and surface contain data that can be loaded into ParaView for visualization — just like the sample animation shown above.