Starting in February 2026, the names of queues are changed to better reflect their hardware. As a result, you may see an error with Error: prompt line 1: in run: Failed to set burst account or queue. Cannot find the queue. See this article for how queue names map to old queue names and AWS Cloud (Amazon EC2) instances.
The Ansys Cloud Burst Compute™ capability is available for the Ansys Lumerical FDTD™ software starting in the 2025 R1.3 release. With this capability, you can seamlessly submit jobs on demand to servers in the cloud equipped with the latest high-performance GPU/CPU hardware managed by Ansys.
Using Ansys cloud Burst Compute™ for Lumerical FDTD does not require any additional setup, and costs pay-as-you-go credits which covers everything required to run the job. You can submit CPU, GPU, and sweeps simulations through the FDTD user interface and scripting, on both Windows and Linux.
Requirements and Onboarding
To use Ansys Cloud Burst Compute from Ansys Lumerical FDTD desktop application, your company must subscribe to an Ansys Cloud plan. Each subscription includes a pool of cloud compute credits and provides access to a limited number of users. On top of your subscription, you can also purchase additional cloud compute credits as needed. Your subscription is managed by administrator within your company using the Ansys Admin Portal. For information on how to use this portal, see the Ansys Admin Portal Guide on Ansys Help.
In addition to the Ansys Cloud Corporate plan and sufficient credits, you will also need a license to enable the Ansys Lumerical FDTD user interface (such as Ansys Lumerical FDTD license, Ansys Lumerical Enterprise or Ansys Lumerical Enterprise prep/post license).
For more information on the Ansys Cloud Corporate plan and Ansys Cloud Burst Compute, please contact us via the landing page for Ansys Cloud Burst Compute or contact your sales representative.
Usage Example – My First Lumerical GPU Simulation on Ansys Cloud Burst Compute
This section demonstrates a simple example of submitting a GPU job using Ansys Cloud Burst Compute with default settings. For more advanced options, refer to the section below on Job Settings and Procedures.
To submit your first simulation job using Ansys Cloud Burst Compute, follow the procedure below:
- Open your project file in Ansys Lumerical FDTD 2025 R1.3 or newer.
- Navigate to the “FDTD” Tab, select “GPU”, and select “Burst” from the “Run Simulation” Group.
- Click “Run”. If you are not logged into your Ansys account, you will be prompted to do so.
- The Job Submission panel opens. Ensure you have sufficient credits and click “Submit” near the bottom of this panel. By default, a queue is automatically selected based on memory estimation of the job. You can select your own queue by checking the “Advanced” box. More information can be found in the “Job Submission Panel” section below.
- Job status and progress are displayed in the Job Manager window. Once it is complete, the completed simulation file containing the results and associated log files are automatically downloaded and loaded into the desktop Lumerical FDTD interface.
Your first simulation with Ansys Cloud Burst Compute for Lumerical FDTD is now complete.
Job Settings and Procedures
This section will outline the detailed options and procedures for job submissions to Ansys Cloud Burst Compute using Ansys Lumerical FDTD.
Sign-In
A sign-in to your Ansys account is required to use Ansys Cloud Burst Compute for Lumerical. This can either be done through the Optics Launcher, through the top right of the GUI, or when you first open the job submission window.
After initiating the sign in through either process, press the “Sign in with Ansys” button, and follow the instructions on the opened webpage.
Submitting Simulations via the GUI
To submit a single simulation job, select “Burst” from the resource selection dropdown menu in the “Run Simulation” Group of the “FDTD” Tab for either CPU or GPU simulations. Pressing the “Run” button will open the Job Submission panel, which will be described in a subsequent section.
To submit a sweep job, go to the Optimization and Sweep window, select a sweep then click the “Burst” button. This button opens the Job Submission panel for the current sweep. You can also submit a sweep job with a right-click on the selected sweep, then selecting “Run Burst” from the context menu. Note that these buttons will be disabled if an unsupported solver (such as RCWA, or STACK) is used for the sweep.
Submitting Simulations via Script
Simulation job settings can also be displayed and set using script commands. Similarly, submission of the job can also be completed via script.
To get the current job settings, you use the getresource script command, that will provide the current settings and available queues. For detailed information on the returned struct including its fields , refer to the getresource script command page.
#Returns a struct with current burst settings
burstSettings = getresource("burst");
#Returns all available burst accounts
burstAccounts = getresource("burst","accounts");
#Returns a list of GPU queues for the Ansys account name Example User
gpuQueues = getresource("burst", "Example User", "GPU");
#Returns a list of CPU queues for the Ansys account name Example User
cpuQueues = getresource("burst", "Example User", "CPU");
To set and submit jobs, the usual run and runsweep commands can be used. Prior to running these commands, you must be logged into your account. When submitting jobs using these commands, the Job Submission panel will not appear. When running these commands without the optional settings as an argument, the default account, settings, and the queue automatically selected by FDTD will be used.
Examples for these commands are shown below. For more information, such as the exact fields of the options struct, see pages on the run and runsweep commands.
#Submit a single simulation burst job with default options using the GPU
run("FDTD", "GPU", "burst");
#Submit a single simulation burst job with the Ansys account name Example User, name it as MyBurstJob, and turn off job monitoring
burstSettings={"account": "Example User", "name": "MyBurstJob", "jobMonitoring": false};
run("FDTD", "GPU", "burst", burstSettings);
#Submit a sweep called “burst sweep” with default options using the GPU
runsweep("burst sweep", "GPU", "burst");
#Submit all sweeps with default settings on the GPU
runsweep("GPU","burst");
You can also use the in-product job manager queue and the script commands addjob and runjobs to submit and run multiple jobs in parallel with Ansys Cloud Burst Compute after setting them up locally.
Note: The command resumejobs is not supported.
When you submit jobs using these commands, you must select a queue using the settings argument. FDTD doesn’t run simulation and memory checks when you submit jobs using this feature. Therefore, when selecting queues, choose a queue that can accommodate the job with the highest memory requirement.
An example for this workflow is shown below, where two jobs are submitted using a queue named “queue_name”.
#Create simulation files named “test1.fsp” and “test2.fsp”
burst_settings.queue = “queue_name”;
addjob(“test1.fsp”);
addjob(“test2.fsp”);
runjobs(“FDTD”,”CPU”,”Burst”,burst_settings);Job Submission Panel
The Job Submission panel displays various options for job submission to Ansys Cloud Burst Compute. An example screen for this panel is shown below, along with a table explaining each field.
| Field | Description |
| Job Name | Name of the job for the Ansys Engineering Portal, you can change the name of the job from the automatically generated default name if needed. |
| Account | The account name used for this job. |
| Account Balance | Remaining balance in your account. |
| Estimated Cost per Hour | An estimation of the cost per hour for your simulation, this value is queue specific. |
| Simulation and Memory Report | Open the simulation memory estimation report for your current simulation. For sweeps, the memory estimation is only for the main (currently opened) simulation file. |
| Download results after completion |
Whether to download results after completion. If this is not checked, you will need to download the result manually through the Ansys Engineering Portal. Cannot be selected if “Submit without monitoring” is checked. This option is enabled by default. |
| Submit without monitoring | Can only be checked if “Download results after completion” is not checked. Enable this option to submit your result to the cloud without any status indication in the GUI. An Ansys Engineering Portal link to the job is provided in the script prompt after submission, and you can continue to work in the GUI without affecting the cloud job. You can use the Ansys Engineering Portal to cancel to job. This option is disabled by default. |
| Advanced | Enable this checkbox to display and select the desired queue for your simulation. If left unchecked, a queue based on the memory estimation will be picked automatically. |
| Queue | Select the desired queue for your simulation. The mapping of queue names to their respective AWS Cloud (Amazon EC2) instances is in this Knowledge Base article. Some queues are marked as Spot instances. A warning on these types of instances is displayed in the submission window. See this help document for more information on Spot instances. Warning: On GPU queues with the “g6e” instance type, you must use and pay for all GPUs on the node. |
Job Manager Status
The job manager opens after you submit a simulation, it contains useful job status information. Three important components for Ansys Cloud Burst Compute for Lumerical are labeled and explained below.
1 – Status
The status box displays the current status of the job, the indicators are as follows:
- Uploading: Uploading file to the Ansys Cloud Burst Compute server.
- Creating: Creating job in the cloud server.
- Queued: Queued for simulation in the cloud server.
- The following status indicates that the simulation is commencing:
- Starting
- Updating Modes
- Meshing
- Building Sources
- Running
- Saving
- Wrapping Up
- Completed: Simulation complete and ready for download.
- Downloading: Downloading simulation from the Ansys Cloud Burst Compute server. This will only appear if “Download results after completion” is checked in the Job Submission panel.
- Downloaded: Simulation completed and results have been downloaded. This will only appear if “Download results after completion” is checked in the Job Submission panel.
- Cancelled: Simulation has been cancelled.
2 – Progress
This progress bar displays the progress for the current state of the simulation, such as uploading and downloading progress.
3 – Overall Progress
This progress bar displays the overall progress of all jobs.
4 – Stop Monitoring button
You can click on the “Stop Monitoring” button to stop monitoring the job and proceed on continuing working in the GUI. A link to the job on Ansys Engineering Portal is provided in the script prompt after you stop monitoring it. If you had selected to download the results after completion in the submission window, it will no longer be downloaded after you stop monitoring, and you must download the completed simulation from the Ansys Engineering Portal manually.
Query Jobs and Download Results via Script
You can also download the results from Ansys Cloud Burst Compute™ with scripts. To see your recently submitted job IDs, use burstrecentids.
#Assume a burst job was submitted with an ID of example-burst-job-id, and that it is the only job
out = burstrecentids;
?out{1};
Returns:
example-burst-job-idTo query the status of a job, use burstjobstatus. This function returns the same status indicators as the job manager status shown above.
#Assume the burst job is finished
?burstjobstatus("example-burst-job-id");
Returns:
FINISHEDAfter a job is complete, use burstresultsquery with that job ID to see the available files.
#GPU burst job with a single simulation and ID of example-burst-job-id
out= burstresultsquery("example-burst-job-id"); #This returns a struct with the job ID and simulation files
sim_results=out.simulation.files; #Obtain the simulation files themselves
?out.simulation.files{1};
Returns:
Struct with fields:
name
size_kB
?out.simulation.files{1}.name;
Returns:
BurstTest.fsp
?out.simulation.files{1}.size_kB;
Returns:
123.45You can then download the simulation results using burstresultsdownload. For sweep jobs, you can download either individual points in the sweep using the ID of the simulation, or the entire sweep using the JobID.
#Assume a GPU burst job with a single simulation and ID of example-burst-job-id
options_struct=struct;
options_struct.dest_dir = "C:\Test_Burst_Download"; #Set a destination directory, the directory must already exist
burstresultsdownload("example-burst-job-id",options_struct); #Download all result from the job
options_struct.dest_dir = "C:\Test_Burst_Download_Logs_Only";
options_struct.file_filter="*.log"; # Download only log files
burstresultsdownload("example-burst-job-id",options_struct); #Download only log results from the job
Ansys Engineering Portal
You can use the Ansys Engineering Portal to view the history of job submissions, monitor the status of jobs, cancel jobs, download simulation files, and monitor credit consumption. For guides on how to do these, see Section 5 of the Ansys Engineering Portal Guide (Ansys account required to view).
Note: When you download GPU simulations from the Ansys Engineering Portal, you must also download the .h5 file and place it in a separate folder with the same name as the simulation file to access simulation results.
Available Features and Current Limitations
There are currently the following limitations for using Ansys Cloud Burst Compute for Lumerical:
- It only supports simulations using the FDTD solver.
- It supports single simulations and sweeps. Optimization jobs are currently not supported.
- It supports single-node multi-GPU up to eight GPUs.
See Also
Introduction to High-Performance Computing with Lumerical, Ansys Cloud Burst Compute