If your laptop is your bottleneck, Inductiva is your breakthrough. Instant, on-demand scale with a pay-as-you-go approach.
This tutorial will show you how to run CM1 simulations using the Inductiva API.
We will cover an example simulation available in the official CM1 v18 release, to help you get started with simulations.
Before running the simulation, you need to prepare the required files and directory structure.
cm1r18 folder, create a new subfolder named input_files.input_files directory. For more information about these input files, refer to Step 2
of the CM1 official guide.| File Path | Description |
|---|---|
cm1r18/run/namelist.input | Simulation configuration file. Mandatory. |
cm1r18/run/LANDUSE.TBL | Defines surface conditions used in radiation and flux schemes. |
cm1r18/run/LANDUSE.TBL | Defines surface conditions used when enabling surface fluxes (heat, momentum, moisture) or the atmospheric radiation scheme. |
cm1r18/src/base.F | Configures base-state conditions. Includes two parts: (1) hydrostatic pressure, temperature, and moisture sounding, and (2) initial winds (u, v). |
cm1r18/src/init_terrain.F | Specifies terrain via the zs array. |
cm1r18/src/init_surface.F | Sets horizontal surface distributions (e.g., heat, moisture). |
Of the files listed above, only namelist.input is required.
The other files are optional and depend on the specific configuration of your simulation. If any of the .F files are not provided, the default files provided by CM1 will be used.
You may also include a custom input_sounding file, although we won’t be using one in this tutorial.
Once all necessary files are in place, your input_files directory should look like this:
-rw-r--r--@ 1 paulobarbosa staff 5125 Jul 26 2015 LANDUSE.TBL
-rw-r--r--@ 1 paulobarbosa staff 63930 Oct 7 2015 base.F
-rw-r--r--@ 1 paulobarbosa staff 53493 Oct 7 2015 init3d.F
-rw-r--r--@ 1 paulobarbosa staff 7952 Aug 31 2015 init_surface.F
-rw-r--r--@ 1 paulobarbosa staff 8952 Aug 13 2015 init_terrain.F
-rw-r--r--@ 1 paulobarbosa staff 6193 Oct 6 2015 namelist.input
Since this is just a introduction tutorial we won't be editing any of the files and will use the default configuration.
Note: In this case, because we are not editing any of the files we could simply just send the file
namelist.inputin theinput_filesdirectory. We added all the other files as an example of what you could do.
Here is the code required to run a CM1 simulation using the Inductiva API:
"""CM1 example."""
import inductiva
# Allocate cloud machine on Google Cloud Platform
cloud_machine = inductiva.resources.MachineGroup( \
provider="GCP",
machine_type="c2d-highcpu-16",
spot=True)
# Initialize the Simulator
cm1 = inductiva.simulators.CM1(
version="18")
# Run simulation with config files in the input directory
task = cm1.run(input_dir="/Path/to/input_files",
sim_config_filename="namelist.input",
# optional config files
base="base.F",
init3d="init3d.F",
init_surface="init_surface.F",
init_terrain="init_terrain.F",
landuse="LANDUSE.TBL",
n_vcpus=1,
on=cloud_machine)
# Wait for the simulation to finish and download the results
task.wait()
cloud_machine.terminate()
task.download_outputs()
task.print_summary()
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.
To adapt the code for this or any other use case, simply replace input_dir with the path to your CM1 input files
and set the sim_config_filename accordingly. Be sure to specify the CM1 version compatible with your input files.
When the simulation is complete, we terminate the machine, download the results and print a summary of the simulation as shown below.
inductiva tasks info 4kemtaacrjjoyr92oksh819my
Task status: Success
Timeline:
Waiting for Input at 13/03, 15:00:21 0.918 s
In Queue at 13/03, 15:00:22 17.618 s
Preparing to Compute at 13/03, 15:00:40 1.823 s
In Progress at 13/03, 15:00:42 364.628 s
├> 1.164 s cp -r /cm1 /workdir/output/artifacts/__cm1
├> 1.151 s cp -f base.F /workdir/output/artifacts/__cm1/src
├> 1.151 s cp -f init3d.F /workdir/output/artifacts/__cm1/src
├> 1.081 s cp -f init_terrain.F /workdir/output/artifacts/__cm1/src
├> 1.066 s cp -f init_surface.F /workdir/output/artifacts/__cm1/src
├> 59.115 s make -C /workdir/output/artifacts/__cm1/src
├> 298.356 s /opt/openmpi/4.1.6/bin/mpirun --use-hwthread-cpus --np 1 /workdir/output/artifacts/__cm1/run/cm1.exe namelist.input
└> 1.065 s rm -r /workdir/output/artifacts/__cm1
Finalizing at 13/03, 15:06:46 2.026 s
Success at 13/03, 15:06:48
Data:
Size of zipped output: 39.11 MB
Size of unzipped output: 128.26 MB
Number of output files: 13
Total estimated cost (US$): 0.0143 US$
Estimated computation cost (US$): 0.0043 US$
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/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 364.6 seconds (around 6 minutes).
It's that simple!
To run your simulation on a larger machine, you’ll need to make a few small changes to both your namelist.input file
and your Python script.
Update the following parameters:
namelist.input:
nodex = 4nodey = 4machine_type = "c3d-highcpu-16"n_vcpus = 16That’s all it takes to scale your simulation to a 16 vCPU machine.
Here are the results of running the same simulation on different machines:
| Machine Type | vCPUs | Execution Time | Estimated Cost (USD) |
|---|---|---|---|
| Local Ryzen 7 7700X | 16 | 1 min, 20s | N/A |
| Cloud c3d-highcpu-16 | 16 | 1 min, 3s | 0.0051 |
| Cloud c3d-highcpu-60 | 60 | 1 min, 25s | 0.014 |
With the Inductiva API, you can easily scale your CM1 simulations to match your computational demands. Whether you need faster runtimes or lower costs, experimenting with different machine configurations allows you to find the optimal balance for your workflow.