13. Frequently asked questions · Troubleshooting
13.1. Purchase
13.1.1. Q. How can I make a payment?
A.
[For corporations]
Payment is due after delivery (post-payment) under the following conditions.
Payment method: Wire transfer or ACH to our designated bank account
Payment due date: End of the month following delivery.
[For individuals]
Pay by credit card via Stripe.
13.1.2. Q. What documents can you issue?
A.
We can issue an invoice, a delivery note, and a quotation.
For Stripe payments, please obtain a receipt from Stripe.
If you need other documents, please contact us. Depending on the request, we may not be able to issue them.
13.1.3. Q. How do I place an order through a reseller or distributor?
A.
Please see Price and Purchase for details.
13.1.4. Q. Can the prices for Winmostar and maintenance be separated? Do you sell maintenance only?
A.
This cannot be split. However, for the educational lineup (V11 and later), you can purchase maintenance separately once the maintenance period for the V11 base license has expired.
13.1.5. Q. Do you provide a warranty?
A.
We do not issue a separate warranty document. The Terms of Use and EULA that you accept when purchasing include terms equivalent to a warranty (including maintenance-related terms). Please use these documents in place of a warranty certificate.
13.1.6. Q. Do you offer discounts for personal use? Do you have an individual license?
A.
No specific restrictions. Individuals may use the Free edition. For the Professional edition, purchase an educational-institution license only if you are affiliated with an educational institution and will use it for institutional work; otherwise, purchase a private-company/public-sector license.
13.1.7. Q. I am affiliated with both an educational institution and a research institute. Can I use my institution’s budget to purchase a license for an educational institution?
A.
No. If the purchasing organization is a research institute, please purchase a private-company/public-sector license.
13.1.8. Q. What is the process for signing a site license agreement when purchasing through a distributor?
A.
As with direct sales, we will conclude a site license agreement between our company and the customer (end user).
13.1.9. Q. Can I purchase on a daily or monthly basis (prorated)?
A.
This is available only in the first year of a site license. For example, if you need a specific-user or node-locked license that works only until the end of the fiscal year due to budget constraints, we can issue a license that is valid only until your requested date (within one year from the license start date). No discount is applied.
13.1.10. Why is Winmostar more cost-effective than similar GUI software products from other companies?
A.
There are several reasons.
One reason is that we do not develop the simulation engines (solvers) in-house; instead, we adopt widely used, well-proven open-source solvers. At the same time, the Winmostar support team understands each solver at the source-code and algorithm level, so we can provide solid support and, when necessary, propose fixes or improvements. This allows you to use Winmostar with confidence in the underlying solver quality.
Another reason is that we separate license fees from support fees, allowing users to choose the level of support they need and reduce costs. Many comparable products bundle support into the license fee, while the support quality is often provided on a best-effort basis (i.e., without guarantee).
Another point is that Winmostar is developed in Japan. For customers purchasing under Japanese domestic terms, pricing is less affected by exchange-rate fluctuations.
Winmostar may appear cost-effective, which can raise concerns about quality. However, Winmostar has been developed and supported for more than 20 years, and that experience is continuously reflected in the product and services. As a result, Winmostar is competitive with higher-priced products in functionality, the volume and quality of documentation (manuals), defect prevention and quality control, and support for beginners.
13.1.11. Q. Is it possible to waive the administrative fee incurred when we are required to use the customer’s procurement system for ordering?
A.
We do not grant fee waivers under any circumstances. The reasons are as follows
13.1.12. Q. When do additional administrative fees apply (e.g., when using procurement systems)?
A.
This fee applies when Winmostar Inc. must log in to customer procurement systems such as SAP or Coupa to submit quotations, place orders, and/or issue invoices. The fee applies even if only one of these tasks is required. It does not apply when using an electronic contract system (see Q. Are there additional administrative fees when using an electronic contract system? for details).
13.1.13. Q. Are there additional administrative fees when using an electronic contract system?
A.
There are no additional administrative fees for signing terms of use and license agreements (site licenses and corporate licenses) using electronic contract systems such as CloudSign, Adobe Sign, and Docusign. Please refer to Q. When do additional administrative fees apply (e.g., when using procurement systems)? for cases where administrative fees apply.
13.1.14. Q. Why is pricing different between Japan and other countries?
A.
Winmostar has a long history of deployment in Japan, and our domestic offering and support processes have been refined through extensive use in the Japanese market. Outside Japan, licensing, contracting, invoicing, and payment processing are handled by Winmostar Inc., together with global service delivery and technical support. International pricing reflects the service scope and operations required to provide stable support to customers worldwide.
13.1.15. Q. Can customers outside Japan purchase at Japanese pricing?
A.
Japanese domestic pricing applies only to customers located in Japan and purchasing under domestic terms. Customers outside Japan are provided licenses under international terms and pricing.
13.1.16. Q. Who issues quotations and invoices for international customers?
A.
Quotations, invoices, and license delivery for customers outside Japan are handled by Winmostar Inc., which manages global sales operations and support for international customers.
13.2. License Code
13.2.1. Q. Can I change the registered user for a specific-user license?
For private companies and public-sector offices, no. For educational institutions, you can change the registered user if at least one year has passed since the last change (or since purchase, if this is the first change). Named User licenses are priced lower because they are restricted to a specific individual.
13.2.2. Q. If I move to another institution, can I continue to use the educational-institution license?
If you are no longer affiliated with the institution at the time of purchase, you cannot use the software—even if it is a perpetual license. This is because educational-institution licenses are specially priced for students, researchers, and faculty who belong to the institution and use the software only for institutional work.
13.2.3. Q. Do I need to reinstall Winmostar when I change editions (e.g., Free → Student, Trial → Professional)?
A. No. Enter the license code you want to use from now on in the menu under Licence code. Your settings will be retained except for the license code.
13.2.4. Q. Can I install multiple editions of Winmostar (e.g., Free and Professional) on the same PC?
A. Yes, it is possible. In this case, please specify a separate installation location for each edition in the Winmostar installer.
13.2.5. How can I change a license code that has already been entered?
Enter the license code you want to use from now on in license code under .
If you are a student and want to switch from the free version to the student version, check the student version on the License registration page and register the license again.
Your settings other than the license code will be retained.
13.2.6. Q. How do I find the MAC address?
A. For Winmostar V11, the MAC addresses of the machine is displayed when it is started with an unregistered license.
Winmostar V10 or lower and Windows 10, first click to launch the command prompt. Next, in the Command Prompt window, type ipconfig /all and press Enter. Various information will be output, look for the “physical address” line in the output. Its contents are the MAC address.
If there is more than one “physical address” line, basically any “physical address” can be used as the MAC address for Winmostar’s node-locked license purchase.
13.2.7. Q. Does the Winmostar license fee include Gaussian and/or VASP licenses? Do you offer a bundled lineup?
A. Winmostar license fee does not include Gaussian and/or VASP license fee. There is no Winmostar lineup with Gaussian license. Please purchase it separately from Gaussianand/or VASP distributor.
13.2.8. Q. What happens when my license expires?
A. When you start Winmostar after the license expiration date, a window for entering the license code will open and the various functions that were previously available will no longer be available. If you update your Winmostar license and enter it again, you will be able to use the software as it was before the license expired.
13.2.9. Q. Can I purchase an educational-institution license for personal learning?
A.
At this time, you cannot purchase a license for an educational institution unless you belong to an educational institution and use it for your own purposes. For individual use, see Q. Do you offer discounts for personal use? Do you have an individual license?.
13.2.10. Q. Q. How do I renew my license?
A.
If you have purchased the Professional Edition, please let us know you would like to renew by replying to the delivery email or via the contact form. If you are using the Free or Student edition, please confirm that you meet the eligibility requirements in the terms of use before registering again.
Once you receive the new license code, overwrite the License Code field with the new code in [Tools] ‣ [Preferences] ‣ [Basic] ‣ [License Code]. After updating the license code, you can continue using Winmostar with your existing settings retained.
13.2.11. Q.. What if a teacher wants to evaluate the Student edition before a supervised student uses it?
A.
Please obtain and evaluate the student version through your current student affiliation. In this case, use for purposes other than evaluation is prohibited. If teachers wish to use the software to further their own research, please purchase the Professional Edition. If you do not currently have any students affiliated with you, please contact us individually via the Contact Form.
13.2.12. Q. What happens when the Professional Edition (Trial) expires? Will I be charged automatically?
A.
After the expiration date, Winmostar cannot be activated. There is no automatic charge for this service. If you wish to continue using the software, please request a quote. Data generated during the trial period will be available for viewing after purchase.
13.3. Support and Maintenance
13.3.1. Q. What is included in maintenance?
It is written in the contents of the Terms of Use. Here it is (Winmostar Terms of Use).
13.3.2. Q. If I have a question, can I email the person in charge directly?
If you have a question, please contact us via the Contact Form rather than emailing an individual staff member. This helps us route your inquiry to the right person and respond faster. If required information is missing from an email, we must ask for it again, which causes unnecessary back-and-forth. Using the form also helps us assign the appropriate person to your inquiry without delay.
13.3.3. Q. What is the difference between paid e-mail support (support on how to operate via e-mail) and free inquiry form questions?
A. Questions or bug reports related to Winmostar-specific functions are covered by maintenance and can be submitted free of charge (please use the inquiry form here: https://winmostar.com/en/support_en.php). For all other matters, support will be provided either as paid email-based operational support or as individual follow-up services. “Winmostar-specific functions” refer to the behavior and operation of the Winmostar GUI itself, and do not include third-party modules (including solvers) that are executed in the background by Winmostar and are developed as open-source or by external parties.
Specifically, for free inquiries, we will focus on the documentation (tutorials and user manuals). If you would like an answer to a question that is covered in the documentation but is more tailored to your research problem, you can ask a question for a fee.
Depending on the customer’s situation, if tailored support is required (such as reviewing calculation conditions or reproducing the issue on our side), the request will be handled as an individual follow-up. If such support is not required, assistance will be provided as email-based operational support.
If you are in doubt, please first use the free contact form to ask your question. We will let you know if we will be able to help you for a fee.
13.3.4. Q. Can I get support for bugs or receive patches?
This will be done in accordance with the terms of use. The latest terms of use are here ( Winmostar V11 Terms of Use ).
13.3.5. Q. How long will you support and maintain previous versions?
Based on the terms of use in effect for your version of Winmostar. In addition, simple instructions on how to operate the system will be provided to the extent possible.
13.3.6. Q. Who develops Winmostar? Where is it manufactured?
X-Ability Co. Ltd.. Place of manufacture is Japan.
13.3.7. Q. Can I contact the developer directly?
You can contact us through the contact form, but our availability is subject to our terms and conditions. The latest terms of use are here ( Winmostar V11 Terms of Use ).
13.3.8. Q. Can I update or upgrade the version of Winmostar I am using?
A. Minor version (and revision) updates can be done as many times as you like during the availability period. Renewal of a major version requires renewal of the license for a permanent license, and is enforceable for an annual license.
As an example, for “V8.039”, “8” refers to the major version and “039” to the minor version. For V9.1.0, “9” refers to the major version, “1” to the minor version, and “0” to the revision.
For example, a perpetual licensee of Winmostar V9 can update from V9.1.0 to V9.1.5 or V9.4.4, but it is not possible to update to V10.0.0.
13.3.9. Q. Are there any precautions when submitting a question?
A. If you have a question that the calculation does not work properly, please send us an input or output file that reproduces the situation in principle.
13.3.10. Q. Do you sell maintenance only?
13.3.11. Q. What happens when the annual maintenance on a perpetual licence expires?
A.
You can continue to use the software in the version that was available before the maintenance expired. Renewing maintenance allows you to update to the version released after the renewal. However, if you wish to upgrade across different major versions, you must purchase a new licence.
13.3.12. Q. Do you provide technical support for the supported solvers? How can I get solver-specific support?
A.
We provide technical support for various solvers, but we do not provide stand-alone technical support for various solvers, except in special cases such as organizational unit partnerships and collaborations.
We provide technical support for various solvers as part of Winmostar ancillary support.
Please purchase Winmostar first and ask email questions within the scope of support that comes with the product itself.
If you have used up the support included with the product, please purchase additional support as needed.
For questions that require the support team to spend some man-hours to answer, we will prepare a separate estimate at that time.
In the case of other vendors, we have seen a form of open source solver technical support that is set at a fixed price regardless of the content, and on the surface unlimited support.
However, in practice, there are often no guarantees as to how much support will be provided for that amount of money, and we believe that such a format is not fair to both the vendor and the user, as there are many variations in the level of difficulty of the support provided. We provide support in the above format.
13.3.13. Q. Do you provide support by phone or via online meetings?
Except for payment-related inquiries, we do not provide support by phone or online meetings. Pre-purchase questions, feature inquiries, and support requests are handled via our contact form or email. This is because (1) Winmostar covers a wide range of methods, and some cases require coordination among multiple support staff, which is inefficient when limited to specific meeting times, and (2) we aim to keep product and support pricing reasonable by minimizing inefficient support. For paid one-on-one follow-up or planning services, we can hold meetings as many times as requested.
13.4. Functions and use of the software
13.4.1. How to pronounce Winmostar.
A. win·mo·str
13.4.2. Q. Does Winmostar connect to internal and external servers or the cloud, or can it be used without a LAN connection?
A. Yes, it is possible. Download the software described in Installing Winmostar and solvers and copy it to a PC not connected to the network. Then, follow the procedure in Installing Winmostar and solvers to install the software.
13.4.3. Q. Please tell me the minimum and recommended specifications of PC for Winmostar.
A. See Hardware requirements for Winmostar itself.
If you are unsure about recommended specs for running solvers (not just Winmostar), consider the following.
This guidance is intended for users who are new to simulations and do not yet have a clear estimate of the calculation conditions or computational load.
In many R&D settings, atomistic simulations are run intermittently rather than continuously like CAE (structural/fluid analysis). Especially at the beginning, simulations are often used on an as-needed basis as a research tool.
At that stage, most time is spent on model setup, parameter selection, and interpretation rather than on raw compute time. You do not need to over-optimize your PC specifications from the start.
In practice, building a solid understanding of the workflow is often more important than the PC’s peak performance.
Poor setup or planning can waste orders of magnitude more time than hardware differences. Focus on establishing a reliable workflow first.
If possible, a multi-core workstation in the JPY 500,000–1,000,000 range is a practical choice, but it is not mandatory.
If your local PC is not sufficient for a specific project, you can temporarily use an HPC or cloud service such as FOCUS supercomputer or HPC Systems Science Cloud to access more compute resources.
After a year or two of use, you will have a clearer view of your typical workloads; choose the next PC based on that experience.
For local jobs, GPUs are not used. Even for remote jobs, GPUs are beneficial only for GPU-enabled solvers and specific workloads. In many cases, CPU resources and memory are the priority.
13.4.4. Q. Is it possible to use the data generated by Winmostar for conference presentations and papers? How should I quote at conference presentations and paper submissions?
A. No problem with using it. When presenting it please quote as About quotation.
13.4.5. Q. Is it possible to upload videos or images of the Winmostar screen to YouTube or other web sites?
A. Yes, you can. When uploading, please cite the URL of Winmostar’s website and state that you are using Winmostar and the version of Winmostar you are using. It is strictly prohibited to upload the preferences window where the license key is displayed.
13.4.6. Q. Does Winmostar include calculation software (solvers) itself?
A. Only MOPAC and CNDO/S are installed in Winmostar. Other software is not bundled with Winmostar due to licensing issues and must be installed separately. Many software can be installed for free, and the instructions are described in Installing Winmostar and solvers.
13.4.7. Q. Does Winmostar run calculations in the cloud?
A. You can run calculations in the cloud, but you can also run them locally. By default, Winmostar does not use the cloud; calculations run on the Windows PC where Winmostar is installed.
13.4.8. Q. Does Winmostar support GPU computation?
A. GPU calculations are supported, but, by default, the GPU is not used. Some solvers are GPU-enabled, but you will have to pay for the work to verify that they work, build them, and configure them. If you have already built a solver with a GPU configuration and have already verified that it works without Winmostar, you can use GPU calculations from Winmostar with only Winmostar configuration. Due to reasons on the solver side, not on the Winmostar side, depending on the combination of OS, machine configuration (including GPU), and solver type/version, it may not be possible to build with a configuration that uses GPU (rather, most cases fall into this category), so before installing a GPU, please contact a vendor that provides hardware for simulation. We recommend that you consult with the vendor that provides the hardware for your simulation.
13.4.9. Q. Which CPU is more suitable, Intel or AMD? Which do you recommend?
A. There are slight differences in performance, but the situation varies depending on the timing. Basically, there is no situation where one is significantly better than the other. Rather than worrying about the difference between Intel and AMD, prioritize stable operation of the computer as a whole and a computer with a large number of cores within your budget. Please contact the hardware vendor for more information.
13.4.10. Q. Can I perform parallel computations using Winmostar?
A. It is possible. Details can be set in the Keyword Settings window for each solver in file mode, or in the Job Settings window in project mode.
13.4.11. Q. Is there an upper limit to the number of cores that can be used for parallel computing using Winmostar? Does the cost vary depending on the number of cores available?
A. You can specify an unlimited number of parallels within the range of user-provided hardware. Winmostar’s price does not change depending on the number of parallels. For local jobs, make sure the requested number of cores does not exceed the MaxCores limit in Winmostar Job Manager. If it does, the job will not start.
13.4.12. Q. Does Winmostar work on macOS and Linux?
A. Winmostar officially supports Windows only. You can check supported Windows versions under System Requirements. To run Winmostar on macOS or Linux, install Windows in a virtual environment such as VirtualBox. Note that some functions (e.g., Gromacs on CygwinWM) may not operate properly, so this is not officially supported.
Linux and macOS can be used as the computer to run the remote job.
13.4.13. Q. Does Winmostar run on ARM-based CPUs?
A. It may work in some cases. As of December 2025, we do not perform monthly automated testing on Arm-based CPUs; however, testing has been conducted for a limited set of configurations, and normal operation has been confirmed (see Hardware requirements for details). As of November 2025, it was confirmed that Gromacs and the solvent placement functions on CygwinWM did not work due to the x86 CPU emulator for Windows 11 on Arm CPUs. Since the x86 CPU emulator for Windows 11 on Arm CPUs is expected to be updated frequently, operation may remain unstable during this period. If you are considering using an Arm-based CPU, please verify functionality in advance using the trial version (https://winmostar.com/en/dlFreeForm.php). We plan to provide regular updates on the status of operation verification on Arm-based CPUs. We are also considering improvements to Winmostar itself and to CygwinWM.
13.4.14. Q. How can I install Gaussian?
For instructions on installing Gaussian, get them from your Gaussian distributor. After installing Gaussian, In the , select the Gaussian program path(g03.exe, g09.exe, g16.exe etc.).
13.4.15. Q. How can I install VASP?
Installation instructions for VASP can be obtained from your VASP distributor. VASP is only available for remote jobs; after installing VASP, edit the template script and specify the directory where you placed the VASP binary and pseudopotential files.
13.4.16. Q. Windows or Explorer freezes when I start the CygwinWM installer. What should I do?
It is very likely that your security software is interfering with the CygwinWM installer, please specify the CygwinWM installer as an exception to be monitored by your security software. This can happen in a variety of cases. This was actually reported on June 9, 2025 when cygwin_wm_20250413.exe was launched in a Windows 11 environment running Trend Micro Apex One.
13.4.17. Q. How many atoms or molecules can I calculate with Winmostar? Is there an upper limit to the number of atoms or molecules that can be handled?
Assuming that the operating speed is not considered, we have confirmed the operation of up to about 1 million atoms. Since the operating speed strongly depends on the execution environment, please check it with the free trial before purchasing. We are planning to increase the speed of Winmostar in future versions.
13.4.18. Q. Can I use Winmostar to calculate coarse-grained models?
As coarse-grained models (i.e., approaches beyond pure quantum chemistry, ab initio calculations, or classical molecular dynamics), Winmostar supports dissipative particle dynamics (DPD) in LAMMPS and the Kremer–Grest model (please inquire separately for Kremer–Grest). United-atom models and related coarse-grained models are planned. For other models, please contact us. In many cases, meaningful coarse-grained modeling requires careful model design and validation rather than “just running the software.”
13.4.19. Q. What is the difference between the latest stable version and the latest development version? Which do you recommend?
A. The latest development version is the latest revision of the latest minor version, and the latest stable version is the latest revision of the previous minor version. (For example, in the case of V11.10.3, 11 is the major version, 10 is the minor version, and 3 is the revision.) When a minor version is updated, new features are added or relatively large specification changes are made. When a revision is updated, defects are fixed and relatively small specification changes are made.
If you do not need to use new features that are only available in the latest development version, we recommend using the latest stable version.
13.4.20. Q. Who owns the rights to the calculation results and output files obtained using Winmostar?
A. Belongs to the user. The user is free to use it.
13.5. Software operations
13.5.1. Q. I cannot create a model as I expected. It cannot calculate. It does not work.
First, please perform the following basic checks.
Check the Installation precautions.
Check whether Winmostar in use is a free version, a student version, a professional version, or a professional version (trial), and whether the function causing the problem is available in that version with function table.
Check the activity record of the security software in use and check whether there is any record in which the activities of the application under Winmostar, CygwinWM and each solver’s installation folder are disturbed.
Update Winmostar to the latest version (it is possible to coexist with the version you are using), check if there is a situation similar to Known problems and Frequently asked questions · Troubleshooting.
If the names of files to be saved and directories containing them (including all upper layers) contain multibyte characters such as Japanese and special symbols including sapces, some solvers may cause problems. Therefore, use only single-byte alphanumeric characters.
Check the working folder to see if the log is being output by the executed process and check the contents of the log.
If calculation is started but the calculation result is wrong, click “Open Log File” etc from the solver menu used in the main menu and check the contents of the log.
For the calculation failure, check that the version of the solver is the same as the version recommended in the Winmostar installation guide. (Especially Quantum ESPRESSO, Gromacs, or LAMMPS)
Disable the “Use Unicode UTF-8 for worldwide language support” option on Windows if it is enabled. (See Some software will not work if the Windows option “Use Unicode UTF-8 for worldwide language support” is enabled. for details)
If a project or file is saved and used on a network drive and unexpected behaviour is observed, save and use it on a non-network drive (e.g. C drive).
Next, please make it possible to record subsequent work with Notepad etc. If you find out how to reproduce the problem, you may be able to fix it in a relatively short time if you report along with the record of work.
Please trace the basics tutorial of the solver you want to use in tutorials.
If you fail to trace the basics tutorial, try the following.
Ensure that the calculations are carried out using exactly the same substances and conditions as in the basic version. (This is not limited to Winmostar, but in all cases where the substances and conditions are different, the calculations may not always end successfully using the methods described in the tutorial).
Trace again to confirm that it is not a misoperation.
If parallel execution is performed, switch to serial execution (parallel number 1).
Restart Winmostar.
Restart the OS.
With security software, Exclude Winmostar, installation folder of CygwinWM, and solvers including MPI from monitoring.
If you are using CygwinWM, run and perform a simple diagnosis of CygwinWM.
Reinstall Winmostar, CygwinWM and the solver you are using.
Try with another PC.
Next, trace the tutorial which seems to be as close as possible to what you want to calculate finally. If it succeeds, please change the calculation condition such as Number of atoms, size of super cell, degree of polymerization, type of element, number of phases, etc. so that it finally gets to calculate what you want to calculate and try to identify where the problem occurs.
Please check for similar cases in Frequently asked questions · Troubleshooting.
If the problem is occured in the part where Winmostar calls external software, Please check also information of the software.
If there is a problem with processing using Cygwin, please confirm Cygwin general problem.
13.5.2. Q. “ERROR: I/O error 32” is displayed and the processing fails.
A. The file related to processing may be opened by an application or process other than Winmostar, locked, or it may have been deleted.
Restart OS and try again in the situation other applications are not open.
13.5.3. Q. Process using Cygwin will end abnormally. / function will display … ERROR …. / Cygwin’s black window shows child_info_fork :: abort: … Loaded to different address: parent …! = Child ….
A. Please do the following procedure one by one from the top, and re-do the process where the error occurred each time.
Perform general dealing with common problems.
Restart your PC
Open Windows Security and go to App & browser control and click Exploit protection. Then change the value of Force randomization for images to Off by default or Use default (Off).
If the value of force image randomization is reset after restarting Windows, start PowerShell with administrator privileges and execute the following command.
cd C:\cygwin_wm Get-ChildItem -Recurse -File -Include *.exe | %{ Set-ProcessMitigation -Name $_.Name -Disable ForceRelocateImages }Then double-click
make_symlink.cmdunderC:\cygwin_wm.
Check the operation with other applications closed. (In rare cases, there may be conflicts with other applications.)
Search and delete anything other than
cygwin1.dlland restart the machineWarning
This operation is necessary in some cases if
cygwin1.dllexists in addition to CygwinWM on the same machine.
cygwin1.dllmay be included in various freeware etc even if you do not install Cygwin in other location.
With all Cygwins on the machine being used terminated, at the [Run] of Windows
C:\cygwin_wm\bin\ash.exe(when CygwinWM is installed inC:\cygwin_wm), execute the /bin/rebaseall -v, and restart the machine.Temporarily disable security software.
Disable software that tends to cause problems described in Cygwin FAQ .
In addition, try the method described in Cygwin FAQ fork() related failures.
Install a new version of Cygwin from the Cygwin official site and check if you can start a terminal.
13.5.4. Q. function displays No reference file (… filelist_cygwinwm.txt) was found….
A. The first thing the CygwinWM diagnostic function does is to look for a file named filelist_cygwinwm.txt directly under the folder where CygwinWM was installed, and the error message is that this file is missing. Please look for filelist_cygwinwm.txt in the location where CygwinWM was installed (by default
C:\cygwin_wm) and specify the folder directly above it in [Tools]-[Preferences]-[Program Path]. Find filelist_cygwinwm.txt in the location where CygwinWM was installed (by default C:\cygwin_wm) and specify the folder directly above it in [Tools]-[Preferences]-[Program Path]. If you cannot find filelist_cygwinwm.txt, the installation of CygwinWM may have failed, so please review the settings of your security software and reinstall CygwinWM.13.5.5. Q. function will display WARNING … some files are missing.
A. Please reinstall CygwinWM.
If it is displayed even after the reinstallation, temporarily disable the security software, or exclude installation destination and installer from monitoring.
13.5.6. Q. function shows ERROR … Failed to generate a current file list.
A. Your Cygwin environment may be corrupted. Please check Cygwin common bugs.
13.5.7. Q. function displays ERROR: Getting SMILES …. OpenBabel process terminates abnormally.
A. If OpenBabel for Windows is installed separately from CygwinWM, Winmostar’s internal processing may use OpenBabel for Windows when it should use OpanBabel for CygwinWM, causing a problem. Uninstall the Windows version of OpenBabel or delete the Windows OpenBabel installation folder from the PATH environment variable.
13.5.8. Q. The job registered in Job Manager is not executed.
If the specified parallel number of MPI is larger than the MaxCore setting of Job Manager, the job will not be executed.
The default value of MaxCore should be set to the number of cores of the running PC, but make sure that it has not been changed or that the number of parallel MPIs has not been set higher.
If you want to run without using Job Manager, uncheck “Run MOPAC using Job Manager” or “Run the other solvers using Job Manager” in Calculation tab of the window.
13.5.9. Q. Even when you stop the job in the Job Manager, it is displayed as ABORT, but the job process continues running.
If the “Default terminal application” in Windows 11 is set to “Windows Terminal”, changing it to “Windows Console Host” may resolve the issue.
13.5.10. Q. When executing a job, an alert saying “Unable to execute (Access denied.)” appears and the job will not start.
A. Please try general errors. In particular, remove the installed Winmostar and Solver folders from the monitoring of your security software.
13.5.11. Q. In the functions where a black console window appears, such as various functions of Winmostar or the execution of the solver, the processing of the black console window does not end and does not proceed.
A. If you happen to click on the black console window, the process will be suspended from there due to the Windows specification.
The processing will resume when ESC key is pressed while the console window is active.
13.5.12. Q. When I open a file or model a molecule, the bonds no longer appear or a lot of extra bonds have come to appear.
A. First of all, in quantum chemical calculations and first-principles calculations (solid-state physics calculations), bond information does not affect the calculation results in any way, so please assume that the presence or absence of bonds is only a display issue.If the bond generation behavior has changed since you installed Winmostar, the bonding factor value in may not be appropriate. Either reset it to the default value or set it to a value around 1.15. If the default value does not give you the desired result, use or to create or delete bonds.
13.5.13. Q. How do I view the extension in Windows Explorer?
- In the case of Windows 7:
Open Explorer
Press the Alt key.
Open View tab in menu
uncheck hide extensions for known files type
- For Windows 8, 10, 11
Open Explorer
Open View tab
Check File name extensions
13.5.14. Q. Markers (red circles) and group selections (blue circles) are not displayed. Atoms can no longer be selected.
Check the checkbox.
13.5.15. Q. The toolbar and tool buttons are not displaying properly. Can I fix it?
First, close Winmostar, make a backup of the UserPref folder, then delete the contents of the UserPref folder and restart Winmostar. If that doesn’t make a difference, try the Winmostar toolbar and tool buttons are displayed incorrectly when Windows display magnification is other than 100%. workaround.
13.5.16. Q. What is an optimization flag?
See Optimization Flags.
13.5.17. Q. Can I have Winmostar not run calculations directly from Winmostar and just save the file, or can I edit the file that Winmostar generates and then run the job?
Yes. Follow the instructions in Basic Operation Flow if you do not want to run the calculation directly from Winmostar and only want to save the file.
13.5.18. Q. Is it possible to use the latest version of each solver or a version not listed in Winmostar’s operating environment?
A. It depends on the type of functionality you want to use with that solver. While we expect most basic functionality to work, we do not guarantee it, so if you want to use a version that is not listed in Winmostar’s Checking, we recommend you check it out in a trial before you buy it. The versions listed in the Confirmation of Operation are the versions used in Winmostar’s pre-shipment testing. For some solvers (Quantum ESPRESSO, LAMMPS, etc.), you can use Tools ‣ Preferences menu to switch the version of the solver that Winmostar outputs.
Winmostar plans to update the supported solver versions as they become available. If there is no specific timeline for updating supported versions and you need immediate support, please consider using paid customization.
13.6. File I/O
13.6.1. Q. I cannot open files created by other than Winmostar with Winmostar. When I edit the file generated by Winmostar and try to open it with Winmostar, it cannot be opened.
A. Please check if the line break code or encoding has been changed. If you still cannot open the file, please attach the file and contact us using the contact form.
13.6.2. Q. When opening a file in MOL or SDF format, the bond length becomes unnatural. Hydrogen does not appear.
Follow the steps below to modify the molecular structure. (1) Adjust All Bond Lengths (2) Regenerate Z-Matrix (3) To marked atoms (Auto) In the case of SDF files, refer to Editing an SDF file procedure.
13.6.3. Q. When I load SMILES, it does not generate the intended structure. Why is this? What should I do?
A. The characteristics of the external program that reads SMILES may not generate the intended structure; switching the external program that reads SMILES may solve this problem. In File -> Import -> SMILES, select either Use OpenBabel or Use Balloon.
13.7. On the modeling of molecules and creation of systems
13.7.1. Q. How do I change the type of chemical bond (single, double etc.)?
A. You can change it by the following method, for example.
1) or by pressing the Add/Change Bond button at the top of the main window multiple times.
2) is selected, the type of chemical bond is automatically changed by the bond degree determined from the interatomic distance. If you optimize the structure with in advance, it may be changed automatically more reasonably.
3) If only one small molecule is displayed, execution of MOPAC calculation will automatically change the coupling order using the population analysis result.
13.7.2. Q. function will display Error: Failed to solvate. and processing will fail.
——— Question details ———
When executing , the following is output to generate.log and processing does not end normally.
gmx insert-molecules -try 100 -f gmx_tmp_water.gro -o gmx_tmp_water_tmp.gro -ci mol0.gro -nmol 64
...
set +v
Error : Failed to solvate.
Try general errors and Cygwin general errors and execute after reducing the number of molecules or reducing the density.
13.7.3. Q. When I run some functions, I get \357\273\277 etc. and the process fails.
——— Question details ———
When I run , the following output appears in packmol.log and the process does not end normally. ::: packmol.log.
/cygdrive/C/winmos11/UserData/builder_tmp/packmol.sh: line 1: $'\357\273\277packmol': command not found
A. \357\273\277 is a code contained in a UTF format text file called BOM and is not supported by programs in Cygwin and Linux environments. If you have checked [Beta: Use Unicode UTF-8 with worldwide language support] in [Clock and Region] - [Change Date, Time, and Numeric Format] - [Administration] tab - [Change System Locale] in the Control Panel, please uncheck the box if you do not mind.
13.8. MPI on local machine, parallel execution
13.8.1. Q. Why does the parallel computation that flows by itself terminate abnormally? The parallel test in the installation test succeeds. Serial calculations also terminate normally.
A. Gradually increase the number of parallels to 2, 4, 6…. and observe the behavior. If the number of parallelisms is too large for the problem size, the program may terminate abnormally. Also, if you are using a generation Intel Core series CPU with a high-performance core (P-core) and a high-efficiency core (E-core), make sure that the number of parallels does not exceed the number of cores in the P-core.
13.8.2. Q. MPICH ends during computation.
——— Question details ———
While executing MPICH, the following error may be displayed and the calculation may be ended prematurely.
op_read error on left context: Error = -1
op_read error on parent context: Error = -1
unable to read the cmd header on the left context, Error = -1
unable to read the cmd header on the parent context, Error = -1
Error posting readv, An existing connection was forcibly closed by the remote host.(10054)
connection to my parent broken, aborting.
state machine failed.
A.
This error is because MPICH uses the network adapter even if it is localonly and the network adapter is disconnected in the middle.
However, if the network adapter has disconnected from the beginning, this error will not occur because MPICH does not use the network adapter.
When computing for a long time using MPICH, please disable the network adapter and start the calculation.
13.8.3. Q. When executing parallel MPI of Quantum ESPRESSO, LAMMPS, warning is displayed as Unable to open the HKEY_LOCAL_MACHINE\SOFTWARE\MPICH\SMPD\process\???? registry key, error 5 , Access denied.
A. MPICH tries to rewrite the registry but it failed because there is no administrator authority.
If you start Winmostar with administrator privileges, you will not get a message. Even when a message is displayed, the calculation itself is executed normally, so it can be ignored.
13.8.4. Q. The installation of .NET Framework 3.5 using dotNetFx35setup.exe fails. What should I do?
A. You may be able to avoid this problem by installing from Windows PowerShell. Start Windows PowerShell with administrator privileges and enter the following command.
Set-ItemProperty -Path 'HKLM:\Software\Policies\Microsoft\Windows\WindowsUpdate\AU' -Name 'UseWUServer' -Value 0;
restart-service wuauserv;
dism /online /Enable-Feature /FeatureName:NetFx3;
Set-ItemProperty -Path 'HKLM:\Software\Policies\Microsoft\Windows\WindowsUpdate\AU' -Name 'UseWUServer' -Value 1;
13.8.5. Q. I cannot find the number of parallels I want to set when I pull down the Number of Parallels combo box. How can I set the number of parallels I want?
Enter a value in the Number of Parallels field using the keyboard instead of selecting it by pull-down. You can enter any value in the Number of Parallels field using the keyboard.
13.8.6. Q. How can I execute multiple jobs sequentially, rather than simultaneously, in local jobs?
A. Set Max Cores in Winmostar JM to the same number as the number of threads or MPI processes used by each job.
13.9. Remote Jobs
13.9.1. Q. How do I submit a job to a supercomputer or a Linux server in my company or my campus?
A. If you need environment settings specific to the computer to which you are connecting, you can do it by creating a template script for remote jobs.
See Remote job for more information.
13.9.2. Q. I cannot connect from Winmostar to a server that allows SSH connections with various terminal software.
A. First check the steps described in Remote job. If this does not solve the problem, you may be able to connect by changing the version of libssh, the library used by Winmostar for SSH connections. Check [Tools] - [Preferences] - [Compute] - [Use older version of libssh2 (1.8.2) for SSH connections].
13.9.3. Q. SSH connection with public key method fails, when I specify the same private key file with other software such as TeraTerm or Cygwin, the connection succeeds. Is there a solution?
A. The key format may not be supported by Winmostar. See SSH public/private key authentication.
13.9.4. Q. The connection test with Test Connection succeeds, but the job submission fails.
A. There are various reasons. Here are some examples
1. If there is a limit on the number of SSH connections such as TSUBAME3.0, please consult us to avoid it without executing SSH connection every time.
2. In some cases, this can be avoided by activating password authentication as well as secret key authentication on the server side.
3. If there are multiple login server entities and they are automatically selected in the background, you may be able to avoid this by using only a specific login server or connecting until all servers are registered in the cache .
4. Even if Winmostar submits a job-submission command (such as qsub) from the local machine, the command may not be found on the remote server. To avoid this, open Profile –> Edit Profile… in the Submit Remote Job window and set Prefix for queueing commands to the directory that contains the command. For example, if qsub is located at /usr/local/bin/qsub, enter “/usr/local/bin/” in Prefix for queueing commands.
13.9.5. Q. Connection test with Test Connection succeeds, but Putty’s WARNING is displayed and various operations fail.
——— Question details ———
Even though TestConnection returns OK, various commands cannot be executed.
And the following dialog is displayed at the time of launching remote job submission window or TestConnection etc.
WARNING: Putty default host name was found in registry.
(\SOFTWARE\SimonTatham\PuTTY\Sessions\Default%20Settings\HostName)
This may cause errors while job submission.
Clear this setting.
A.
Reason:
This WARNING occurs when Putty's HostName is set.
Since Putty's settings are saved in the Windows registry, even if the Putty is other than Winmostar bundled version, this problem will occur when some sort of character string is saved in HostName.
Solution:
Launch Putty from Remote Job Submit window. Check a character string is set in the HostName field of Default Settings.
You can solve this problem by deleting this string and saving it with Default Settings selected.
(The input contents in the Port column are not particularly affected.)
13.9.6. Q. What type of MPI (MPICH, OpenMPI, etc.) can I use on the remote server?
A. Basically any type of MPI is available. We have experience with MPICH, OpenMPI and MVAPICH.
By editing the template script, you can freely execute commands such as source, module, and export, and set up the environment for executing any MPI.
The solver used must be compiled with the MPI (mpicc, mpif90) used.
13.9.7. Q. Why are some environment variables (e.g. PATH) that should be set on the remote server by filling in bashrc, etc., not set when running from Winmostar? What should I do in such cases?
A. This is because the processing flow on the remote server may differ between interactive shell operations in the terminal and direct command execution with ssh command arguments. bash command arguments may do the same. As a workaround, please create a template script and explicitly fill in the environment variable settings, etc. separately in the script.
If you have an Ubuntu machine including WSL2 as a remote server, please also refer to here
13.9.8. Q. Is there any difference in behavior between the Windows version of the solver and the Linux version of the solver?
A. There is essentially no difference in functionality except for calculation speed. Performance varies depending on compiler, MPI, and compilation settings.
13.9.9. Q. Can I run calculations simultaneously for remote jobs that run on different servers?
A. Yes. However, in project mode, each job is tied to the runtime profile information, so create a different profile for each server.
13.9.10. Q. Where are passwords entered in the remote job profile stored? Is the password information stored in the project folder?
A. Passwords are stored encrypted in winmos_profile.ini under the user settings folder (UserPref in the Winmostar installation folder). It is not stored anywhere else, so there is no password information stored in the project folder. For example, when you send us your project folder for support, we will not send you the password information.
13.9.11. Q. Why can I SSH to an Ubuntu machine, including via WSL2, but cannot successfully execute a remote job?
A. On Ubuntu machines, because of the following part written at the beginning of the .bashrc file, remote job execution does not reflect the PATH and other settings that are set after this.
# If not running interactively, don’t do anything
case $- in
*i*) ;;
*) return;;
esac
Delete or comment out these lines.
Also, for WSL2 Ubuntu, since zip is not included by default, please also run the following command to install zip.
sudo apt-get install zip unzip
13.9.12. Q. Job submissions have suddenly started failing on external remote servers such as the FOCUS Supercomputer. What could be the cause?
A. You may have exhausted your allocated CPU usage quota, or your disk usage may have reached its limit. Please check the resource usage status on your system.
On the FOCUS Supercomputer, starting in October 2026, the free home1 storage quota per project has been reduced from 200 GB to 10 GB. If your project ID is xxx, please check your disk usage by entering: df -h /home1/gxxx To run this command, select your remote server profile under , and then enter the command in the white area of the Control Remote Job/Server window.
13.9.13. Q. How can I delete files on an external remote server such as the FOCUS Supercomputer?
A. In Project Mode, you can delete an entire project or everything under a specific working folder at once. Right-click the project or working folder you want to delete in the main window, and then select Delete.
If you wish to delete files manually, please log in using tools such as PuTTY or WinSCP and perform the deletion directly on the remote server.
When deleting files on the FOCUS Supercomputer, please make sure you are connected to the VPN in both cases before performing any operations.
13.10. Simulation in general
13.10.1. Q. Is it possible to calculate the interaction between one substance and another substance with Winmostar?
A. This kind of question is very common, but the definition of the word “interaction” is broad, so it depends on the definition.
First of all, please check the manual of each solver beforehand, because Winmostar depends on what each solver (Quantum ESPRESSO, GAMESS, Gaussian, LAMMPS, Gromacs) can do, in terms of whether it can perform quantum chemistry, ab initio calculations, and molecular dynamics itself for the material you are focusing on.
Secondly, interactions in the sense of wanting to know the mechanism of some physical property or phenomenon that we are focusing on cannot be answered immediately, because it requires knowledge specific to that material system, and the investigation of whether it is computable requires sufficient investigation itself. We also offer a service to conduct investigations for a fee for private companies
If you want to quantitatively evaluate the bonding or association caused by the interaction, you can discuss the change of energy versus bonding distance for absolute zero, or the radial distribution function between binding materials or the autocorrelation function of the bonding time series data for the kinetic state.
Finally, if you know the specific analysis, such as Kitaura-Morokuma analysis or interaction energy analysis such as Coulomb-vdW interaction in MD, please ask about it.
13.10.2. Q. Can I calculate bond energies and bond divergence energies for a given bond in a molecule?
A. Perform quantum chemical calculations of the stable conformations of the bound and dissociated states and calculate from the energy difference between them. The two molecules in the dissociated state should be calculated separately, and the spin and charge should be set carefully.
13.11. MOPAC, CNDO/S, GAMESS, NWChem, Gaussian
13.11.1. Q. I get an error while running a quantum chemistry calculation, what should I do?
A. First trace the Winmostar tutorial. If that fails, you may have failed to build the environment, and for Gaussian, also check if the license allows parallel computation.
Next, if you did the calculation in project mode, check the error column in the working folder for the corresponding calculation. You can also hover the mouse pointer over the working folder to see the error. If the error message is in the FAQ, try the workaround described there. Especially for IR calculations of large molecules, the default amount of memory specified may not be enough and the calculation may stop.
If the energy (SCF) calculation does not converge, please check or try the following.
- Check for proper initial structure, charge, and spin multiplicity.
- Using large basis functions and diffuse functions makes convergence difficult. Consider using smaller basis functions.
- If the DFT calculation in GAMESS does not show “DFT CODE IS SWITCHING BACK TO THE FINE GRID” after displaying “CONVERGED TO SWOFF, SO DFT CALCULATION IS NOW SWITCHED ON” without approaching convergence please set the variables NRAD0 and NLEB0 in the DFT tab of the Keyword Setup window.
- In Gaussian, if the result of a small basis function is used as the initial trajectory, it is more likely to converge in the large basis function. In project mode, the energy calculation is first performed on the small basis function, and then a continuation job is performed on the large basis function. At that time, the continuation source should be specified as a small basis function calculation, and in addition, the guess=read checkbox should be checked in the Keyword Setup window.
- For Gaussian and GAMESS, running two or more local jobs simultaneously may cause abnormal termination due to memory management and related issues. This can be resolved by executing the jobs sequentially. Please refer to Q. How can I execute multiple jobs sequentially, rather than simultaneously, in local jobs? for instructions on how to run jobs sequentially.
13.11.2. Q. The Hamiltonian calculation method (Hamiltonian) or basis function (Basis Set) that I want to use is not found in the Configure window or the Easy Setup window for MOPAC, GAMESS, Gaussian and NWChem. How do I set it up?
A. If you can enter directly on the keyboard in the Calculation Method and Basis Function setting fields, you can do so directly. Please check the manual of the respective solver, as the description of how to specify polarization functions (represented by “*”, “p”, “d”, or “f”) may differ from solver to solver.
13.11.3. Q. How are bonds determined when opening an input/output file or when regenerating bonds?
13.11.4. Q. Unnecessary bonds are displayed while modeling the system or when reading the calculation results. Or the bonds are not displayed. How can I fix this?
A. See this FAQ for more information.
13.11.5. Q. Is it possible to optimize substructures from Winmostar using quantum chemical calculations?
A. As of V11.15.0, the support is as follows
MOPAC6: Only Z-Matrix is supported
GAMESS is supported
Gaussian is supported
NWChem: only Z-Matrix supported
13.11.6. Q. Why does the Dipole moment value displayed at the bottom of the Winmostar molecule display area differ from the dipole moment value manually calculated from the charges assigned to each atom, such as Mulliken charges?
A. The Dipole moment value displayed at the bottom of the Winmostar molecule display area is the dipole moment for the entire molecular system directly output by MOPAC, GAMESS, Gaussian, and NWChem. This value is the exact dipole moment calculated by calculating the center of gravity of all negatively charged electrons and the center of gravity of all positively charged nuclei. In the calculation of this value, no operation to assign an electron to each atom is involved.
In atom-by-atom charge schemes such as Mulliken charges, the boundary between atoms is not uniquely defined and remains somewhat arbitrary. In addition, the charge-assignment formula differs by method (e.g., Mulliken vs. NBO). Therefore, dipole moments computed from per-atom charges may differ slightly from the total dipole moment for the entire molecular system reported by MOPAC, GAMESS, Gaussian, or NWChem and shown at the bottom of the Winmostar molecular display area.
13.11.7. Q. How many atoms can be calculated with MOPAC?
A. 70 heavy atoms (other than hydrogen) and 90 light atoms (hydrogen).
You can also download the executable binary of MOPAC6 for large molecules (up to 420 atoms) from the here <https://winmostar.com/jp/mop6wxxx.zip>.
Winmostar has been confirmed to work with MOPAC opensource version.
MOPAC opensource version has been available free of charge from 2021 with no limit on the number of atoms.
MOPAC Page <https://openmopac.net/>`_
13.11.8. Q. It will display abnormally with the message ATOMS **, **, AND ** ARE WITHIN. **** ANGSTROMS OF A STRAIGHT LINE displayed in the MOPAC log.
——— Question details ———
It stops with an error saying that 3 atoms are linear as shown below.
CALCULATION ABANDONED AT THIS POINT
THREE ATOMS BEING USED TO DEFINE THE
COORDINATES OF A FOURTH ATOM, WHOSE BOND-ANGLE IS
NOT ZERO OR 180 DEGREEES, ARE IN AN ALMOST STRAIGHT
LINE. THERE IS A HIGH PROBABILITY THAT THE
COORDINATES OF THE ATOM WILL BE INCORRECT.
THE FAULTY ATOM IS ATOM NUMBER 69
Finally, comes
ATOMS 68, 57, AND 54 ARE WITHIN .0134 ANGSTROMS OF A STRAIGHT LINE
.
A.
It is displayed when an angle close to 180 degrees is included in Z-Matrix.
Change the connectivity of atoms by the coordinate editing function in the lower right of the main window so that there is no angle close to 180 degrees from Z-Matrix.
If you are new to Z-Matrix, you may be able to avoid this error by adding “XYZ” to the keyword as a different method.
Alternatively, by adding a dummy atom of the atomic type XX to a position out of the line where three atoms are aligned in a straight line and designating it as a connection destination on the Z-Matrix of the atoms arranged in the line,
Sometimes it is possible to avoid errors.
13.11.9. Q. When MOPAC is executed, “UNRECOGNIZED KEY - WORDS: (PM 6 (Hamiltonian name))” is output to the log and the calculation ends.
A. If you can avoid the error by changing Hamiltonian to AM1 with MOPAC Configure, the selected Hamiltonian is not supported by the MOPAC you are using.
Check the list of Hamiltonians supported by each version and select the appropriate Hamiltonian.
If it still does not work, please take action of general errors .
13.11.10. Q. How do I use the solvent effect (COSMO method) in the MOPAC attached with Winmostar?
MOPAC which is included in Winmostar does not support the solvent effect (COSMO method). Please consider using GAMESS which is more accurate than MOPAC, although it takes more time to calculate.
13.11.11. Q. When I calculate with MOPAC other than Winmostar built-in (e.g. MOPAC2016), only some molecular orbitals are displayed.
Check the ALLVECS checkbox in the MOPAC Setup window to create an input file and perform the calculation.
13.11.12. Q. It cannot be calculated when Method=INDO is used in CNDO/S.
A. The elements after F are not supported by Method=INDO.
set Method=CNDO, or use a non-empirical solver such as GAMESS.
13.11.13. Q. Is there a guideline for the maximum number of atoms (memory requirement) that can be calculated in GAMESS, Gaussian, and NWChem?
A. There is no simple guideline for the amount of memory required, as it depends on the molecule size, basis functions, calculation contents, and number of parallels.
When a calculation is actually performed and stops due to insufficient memory, the log file will indicate how much is missing, so it is common to make a judgment based on that value.
13.11.14. Q. Is it possible to speed up 3D display of ESP (electrostatic potential) of GAMESS, Gaussian, and NWChem?
A. If you have installed the Windows version of Gaussian, check the Cubegen check box in the Cubegen window that appears when you open the Cube file, and the processing will be relatively fast using the Cubegen program included with Gaussian.
In the future, we plan to speed up the cube file processing program (OpenCubegen) included with Winmostar.
13.11.15. Q. When running GAMESS, “* ERROR: MEMORY REQUEST EXCEEDS AVAILABLE MEMORY” etc. will be output to the log and the calculation will end abnormally.
A. This means that the amount of memory specified in the input file is not enough.
Increase the value of MWORDS in the $SYSTEM column of the Advanced tab of the GAMESS Configure window. (MWORDS number) x 8MB is the maximum amount of memory used per CPU core. For example, if MWORDS is 200, the maximum amount of memory per CPU core is 200 x 8MB = 1600MB = 1.6GB. Please set MWORDS according to your computer and the number of parallels.
13.11.16. Q. Immediately after installing GAMESS for Windows, I was able to use it without any problems, but at some point, all the calculations started to terminate abnormally in the middle. Is there a way to fix this?
A. First of all, please take action of general errors. If this does not work, perform the following steps in order.
1. Launch the GAMESS installer -> Select Remove -> Uninstall GAMESS.
2. Go to Windows Settings -> Apps -> Apps and Features -> Select Microsoft MPI -> Uninstall Microsoft MPI
3. Reinstall GAMESS and MS-MPI according to the GAMESS installation manual for Windows in the Winmostar Installation Guide.
In particular, when it shows “mpiexec … server rejected credentials” etc. in the console window when running GAMESS, this method may be effective.
13.11.17. Q. GAMESS outputs “WARNING. NUMBER OF INTERNAL COORDINATES IS GREATER THAN (3N - 6), BUT NO SYMMETRY COORDINATES ARE GIVEN.” in the calculation of only one atom and abnormally ends.
A. This is a message indicating a problem caused by using Z-Matrix in a system with only one atom.
In this case, Cartesian coordinates are used (COORD=UNIQUE) to solve it.
Set UNIQUE to COORD on Wimostar's GAMESS Configure window.
13.11.18. Q. The message **** ERROR **** PCM SPHERE (S) MUST HAVE A POSITIVE RADIUS is displayed in the log of GAMESS and abnormally ends.
A. GAMESS may not contain Cavity radius for some atoms.
To specify the Cavity radius, add the following statement just after the $PCM line.
$PCMCAV RIN(13)=1.55, RIN(15)=1.55 $END
In this example, Cavity radius is set to the 13th and 15th atoms.
13.11.19. When running GAMESS, the following messages are output to the log: “ERROR: BAD DELOCALIZED COORDINATES GENERATED!!!” and “THERE ARE NOT ENOUGH PRIMITIVE INTERNAL COORDINATES” causing the calculation to terminate abnormally.
A. In order to reduce the number of structural optimization cycles, Winmostar normally specifies the use of internal coordinates such as bonds and angles, but due to reasons such as the presence of multiple molecules such as dimers, or an excessive number of bonds from a specific atom, the conversion from xyz coordinates to internal coordinates fails and the process stops.
In the GAMESS keyword setting window of Wimostar, select the Z-Matrix tab –> Uncheck $ZMAT and perform calculations using xyz coordinates.
13.11.20. Q. When I try to run NMR calculation in GAMESS (using RUNTYP=NMR), I get an error.
A. GAMESS NMR calculations are only compatible with the closed-shell Hartree-Fock method, and other DFT methods cannot be used.
Also, the execution of NMR calculations in GAMESS can be slow, so we recommend using NWChem or Gaussian instead.
If you really want to calculate with GAMESS from Winmostar, uncheck the DIRSCF checkbox in the $SCF of the Advanced tab in the Setup window.
In addition, parallel computing is not supported, so please set the parallel number to 1.
You can find detailed instructions in the log file with an error, so please refer to that as well.
13.11.21. Q. When I use Diffuse functions in GAMESS, the SCF calculation does not converge.
A. Add ICUT=11 to the Others column in the $CONTRL box on the Basic tab of the GAMESS Setup to make the cutoff value for the two-electron integration smaller (stricter).
13.11.22. Q. For some molecules, after optimization of Firefly, when I check the Molecular Orbit UVvis…, I get the error '*******' is not valid floating point value”.
A. The linear dependence of the basis is large because 6-31+G* has the diffuse functions.
Therefore, a part of the value of the molecular orbital coefficient becomes very large, and **** is output in the log.
The solutions are
1. Use 6-31G* basis set
2. If you use 6-31+G*, use GAMESS instead of Firefly
.
Because GAMESS has processing of linear dependency,
Firefly and GAMESS may have slightly different energy values.
execute a series of calculations with either Firefly or GAMESS.
13.11.23. Q. How do I deal with the message “THE GEOMETRY SEARCH IS NOT CONVERGED!” when I run a structural optimization calculation with GAMESS?
If the animation shows structural optimization information and the energy continues to drop, the calculation has been completed with insufficient number of structural optimization cycles. Please perform the structural optimization calculation again with the final structure.
For some larger molecules (roughly 100 atoms or more), structure optimization with DLC=.T. may result in a significant loss of structure. In that case, unchecking $MAT’s Set DLC=.T…. unchecked and run, which may resolve the problem.
13.11.24. Q. I cannot download the GAMESS installer. How can I do this?
A. If an error occurs on the download site and the registration form does not appear, please try another browser. If you are using Microsoft Edge, try accessing the site again in Internet Explorer mode. Also, sometimes the download site system is down, in which case please wait up to a week.
13.11.25. Q. The DFT calculation in GAMESS does not converge after displaying “CONVERGED TO SWOFF, SO DFT CALCULATION IS NOW SWITCHED ON”. Is there any way to deal with this?
A. GAMESS DFT energy calculations are typically performed in three steps to improve convergence and calculation efficiency. The first step is a Hartree-Fock calculation, followed by a DFT calculation on a rough grid, and finally a DFT calculation on the default or input-specified grid. When “CONVERGED TO SWOFF, SO DFT CALCULATION IS NOW SWITCHED ON” is displayed, the second step of DFT calculation with the rough grid starts, but for some molecules, the grid is too rough to converge and the “DFT CODE IS SWITCHING BACK TO THE FINE GRID” may not be displayed.
To increase the number of grid points in the second step, enter “NRAD0=99 NLEB0=302” in the Others field on the DFT tab of the GAMESS Keyword Setup window.
13.11.26. Q. When running NWChem in parallel, the calculation outputs “Please specify an authentication passphrase for smpd:” to the log and does not run.
A. If you omit the pass phrase (passphrase) when installing MPICH 2, such an error may occur.
There are several solutions, but after uninstalling MPICH2, reinstalling MPICH 2 may resolve it.
In that case, you need to stop smpd before uninstalling MPICH2 and install smpd after reinstalling MPICH2.
Also, if you are using 32-bit MPICH2 on 64-bit Windows, use the 64-bit version.
13.11.27. Q. When I run NWChem, I get the message “sym_geom_project: sym_center_map is inconsistent with requested accuracy” in the log and the calculation does not run.
A. If the structure deviates slightly from NWChem’s molecular symmetry criterion, it will stop with an error. Select [Tools]-[Point Group Analysis] in the Winmostar main window and click [Analyze], then [Symmetrize] to make the structure with exact symmetry, or check noautosym in the Advanced tab of the Keyword Settings window, or move one atom slightly to make the structure with broken symmetry.
13.11.28. Q. How can I put the solvent effect in NWChem?
A. Please put the name of the solvent molecule in ***. Please see https://nwchemgit.github.io/COSMO-Solvation-Model.html#solvents-list-solvent-keyword for more information about supported solvent molecules.
13.11.29. Q. When executing NWChem, lines including “out of memory”, “alloc failed”, “ma_push failed”, “nga_create failed”, and “Global memory limit unreasonable” are printed in the log and calculations do not run.
A. None of the calculations are stopped because the amount of memory specified in the input is not sufficient. Please increase the value of Memory in the Basic tab of the NWChem Keyword Setup window.
13.11.30. Q. The Gaussian calculation may or may not work correctly.
A. In case of single core version of Gaussian, Gaussian will be terminated when you try to run more than two jobs. If you want to run two or more Gaussian local jobs sequentially on Winmostar using a single-core version of Gaussian, set the parallel number of each job (%nproc or %nprocshared) to 1, and set Max Cores of Winmostar JM to 1. Please set the Max Cores of Winmostar JM to 1.
In addition, even when using a non-single-core version, particularly for Gaussian09 (g09) or earlier on Windows, running multiple Gaussian jobs simultaneously may result in abnormal termination due to memory management and related issues on the Gaussian side. This can be resolved by executing the jobs sequentially. Please refer to Q. How can I execute multiple jobs sequentially, rather than simultaneously, in local jobs? for instructions on how to run jobs sequentially.
13.11.31. Q. I read the Gaussian log file, but the orbital (eigen) energy etc. are not written.
A. It is not output if pop=full and gfprint is missing in the Gaussian input file.
When you open a Gaussian input or output file in Winmostar, not only the structure but also the calculation conditions will be loaded. If you open such a file and set the keywords, make sure that pop=full and gfprint are set.
13.11.32. Q. How do I calculate the chk file read in Gaussian?
A. For remote jobs, check [Advance] in the SubmitJob window and uncheck [Delete *.chk] to leave the chk file. And the chk file is loaded and the calculation runs when the same name was used.
The method using –Link1– is simpler in setting itself, so please consider using it as well.
13.11.33. Q. I get an error when I try to load a Gaussian fchk file that is several GB in size and display molecular orbitals.
A. We can handle fchk files up to roughly 2 GB. The upper size depends on the molecule size and basis functions. We plan to eliminate the fchk file size limitation in the future.
13.11.34. Q. How can I perform vibration calculations in Gaussian that take isotope effects into account?
A. Set freq=readisotopes in the Freq column of the Gaussian Keyword Setup window and write the following items one line at a time in the Subsection column.
Temperature (K) Pressure (atm)
Mass of the first atom
Mass of the second atom
…
Mass of the n-th atom
13.11.35. Q. When running a structural optimization calculation in GAMESS, the log outputs “OPTIMIZATION ABORTED. – GRADIENT OUT OF RANGE – MAXIMUM ALLOWED FORCE (FMAXT)” and the calculation terminates abnormally.
A. The differential value of energy with respect to the atomic nucleus coordinates (the force acting on the atom) is too large, causing the calculation to halt. There may be an issue with the initial structure or calculation conditions, such as charge or spin multiplicity. Please verify that these input values are appropriate.
13.12. LAMMPS, Gromacs
13.12.1. Q. I would like to run molecular dynamics calculations, but which should I use, LAMMPS or Gromacs?
A. If you are not sure, we recommend LAMMPS, which is slower than Gromacs, but it supports a wider range of force fields and allows for more flexible schemes for non-equilibrium calculations. Winmostar allows you to use LAMMPS and Gromacs with the same ease of use, making the transition to Gromacs relatively easy.
13.12.2. Q. I get an error while running a molecular dynamics calculation, what should I do?
A. First trace the Winmostar tutorial. If that fails, you may have failed to build the environment.
Next, check if the error is reproducible. If it is not reproducible, it is often an environmental (either or both hardware and software) problem or a random number derived problem.
Next, calculate one molecule of each component, without charge. If you still get an error, change the type of force field as well.
If no error occurs with one molecule of each component and no charge, set the charge or increase the number of molecules and observe the behavior.
Next, if you get errors with long calculations, increase the equilibration time.
If you get an error during the equilibration process, try the following.
- Check for conservation of conserved quantities (total energy, extended Hamiltonian) and reduce the time increments.(1 fs→0.5 fs→0.2 fs→0.1 fs, and so on, gradually decreasing the time increments to find the value at which stable operation is achieved.)
- If parallel computation is not performed or it is impractical to set the number of parallels to 1 due to computation time issues, the number of parallels should be reduced gradually.
- If parallel computation is not performed or it is impractical to set the number of parallels to 1 due to computation time issues, the number of parallels should be reduced gradually.
- Use Berendsen and velocity scaling methods for temperature and pressure control.
- Use double precision (for Gromacs)
- Do not use constraints (such as the SHAKE method). Instead, use time increments of about 0.1 fs.
- If the cell size is less than twice the cutoff radius, the calculation fails due to the implementation of the solver, so the number of numerators is increased to increase the cell size and the calculation is performed again.
- Separate the calculation by a short number of steps and repeat the continuation. (Especially useful when initial and final densities differ significantly during pressure control).
- Change the initial density (decrease or increase).
Change the random number seed for initial speed generation.
Change the random number seed for initial structure generation.
If it finishes successfully up to the NVT calculation but fails in the NPT calculation, see Q. When pressure control (constant NPT or NPH) is used, the calculation fails in the middle of the process. Is there a solution?.
13.12.3. Q. How can we equilibrate the system in molecular dynamics calculation?
A. First of all, describing the case where the calculation of the aggregation system (liquid and solid, not gas) of the equilibrium state of low molecular is target.
When aligning molecules in the initial state, please set the density as close as possible to the final density.
However, when it is not possible to arrange it unless it is considerably low density, it does not matter.
After that, please carry out energy minimization, constant temperature calculation, constant temperature pressure calculation until the change of the potential energy, temperature and density converge.
If the initial density is too low, compress it with a pressure higher than the target pressure (for example, about 100 times) by constant temperature pressure calculation.
Finally, if you are interested in the ensemble average physical quantity and reaching the target temperature and pressure after equilibration, the difference in the fine equilibration procedure is less likely to have a significant influence on the calculation result.
For polymers and glasses, it is often impractical to reach a strictly equilibrated state within a realistic computation time. In practice, you run the equilibration long enough for key observables (energy, temperature, density, etc.) to stabilize and for the autocorrelation of the property of interest to decay sufficiently (i.e., the system “forgets” its initial state).
Since pressure control is unstable in the case of gas, equilibrium state is obtained only by energy minimization and constant temperature calculation.
Additionally, for molecules with slow relaxation, such as rigid or bulky molecules, equilibrium density may be difficult to reach. In such cases, please refer to Q. The final density obtained from the NPT calculation is much lower than the density predicted from experimental values and other sources. What should I do?. If equilibrium density cannot be reached within a computationally feasible timeframe, multiple calculations with different initial configurations may be performed and the results averaged (e.g., calculations of solvent-free energy for small molecules in polymer melts may average results from tens to hundreds of different initial configurations).
13.12.4. Q. I ran MD calculations from the same initial configuration but got different results. Why is this?
Floating-point arithmetic is not associative: the result can change slightly when the order of additions and subtractions changes (typically in the least significant digits). In parallel runs, the reduction order can vary depending on communication timing, so parallel calculations are generally not bitwise reproducible. Molecular dynamics is also chaotic: tiny numerical differences can grow exponentially over time and lead to different atomic trajectories. This behavior is expected. With sufficient equilibration and sampling, statistical properties (ensemble averages, distributions) should converge to similar results even if individual trajectories differ.
13.12.5. Q. The final density obtained from the NPT calculation is much lower than the density predicted from experimental values and other sources. What should I do?
First, check to see if it has equilibrated sufficiently, and if it is the last case where the temperature, potential energy, pressure, and density are still changing significantly, calculate for a longer simulation time.
Second, if the molecule is rigid or bulky (e.g., with multiple aromatic rings), it may take a very long time to converge the density to the equilibrium value at the target temperature and pressure—especially for liquid crystals and large monomeric polymers. Because such molecules aggregate easily and reorganize slowly, density equilibration can stall. In such cases, you can accelerate equilibration by intermittently running at a higher temperature than the target (e.g., 600 K when targeting 300 K) and then returning to the target conditions.
13.12.6. Q. In the MD calculation, how does the constraints by the SHAKE method etc. affect the calculation result? How do I choose a constraint method ?
A. The SHAKE, RATTLE, LINCS, and SETTLE methods are applied between atoms that are covalently bonded to restrict the bond length, so that the time step can be increased and even longer computations can be performed more stably with the same amount of computation. “Stable” is in terms of Hamiltonian (total energy) conservation.
Functions that express covalent bonds when not constrained also do not accurately represent actual phenomena. Therefore, assuming that stable calculations are being performed, there are deviations from the actual phenomena due to the respective circumstances both in the case of restraining and in the case of not restricting, in terms of the effect on various calculated physical properties.
Unless the vibrational motion in the molecule itself has the purpose of calculation, it is basically recommended to select the condition that the Hamiltonian preserves for a long time in each case.
However, if the bonds of the hydrogen atoms are not constrained, the atoms will move at high speed significantly in the system, which may cause the drift of the Hamiltonian. Therefore, in many cases, the bonds of the hydrogen atoms are constrained.
13.12.7. Q. Why do molecules sometimes come out of the cell when in the trajectory or the final structure afterMD calculation?
A. Using periodic boundaries, the entity of the molecule should be inside the cells of the periodic boundaries.
But, solvers such as Gromacs and LAMMPS record trajectories with translational values as they are without folding coordinates even when molecules cross over cell boundaries to calculate mean square displacement, etc.
Either way, the result will be properly considered and the same result will be output during the result analysis, so there is no effect on the result analysis.
If the appearance of molecules jumping out of the cell is not good, adjust the settings of View - Wrap Around Cell Boundary.
13.12.8. Q. When I create a cell by adding water molecules from a cell in the [Solvate/Build Cell] menu to a large molecule such as a protein, the molecules sometimes stick out of the cell.
A. Select [View]-[Wrap/Unwrap Around Cell Boundary]-[Wrap for each atoms] and if there is no interference between protein and water molecules, then there is no problem in the calculation.
If you want to center the protein, which is a matter of appearance only, select the atoms near the center before adding the solvent and perform [Edit]-[Adjust Axes]-[Set Origin to Marked Atom] to move the molecule near the center.
13.12.9. Q. The [MD]-[Solvate/Build Cell] and [Insert Molecule] processes do not finish after many minutes. Is there a solution?
A. Reduce the density. When the molecules are relatively small (generally within 30-40 atoms), set the density at 50-75% of the expected density. If the molecules are large or depending on the shape of the molecules, it may be necessary to further reduce the density (several percent to 10%, etc.), so observe the behavior while gradually reducing the density. Even if the initial density is far from the original density, it is expected to eventually change to an appropriate density by running MD calculations with pressure control.
13.12.10. Q. In the [MD]-[Assign Charges Automaticcally] menu, the error message “Topology file not found” is displayed. Is there a solution?
It is very likely that the molecule you are trying to assign a charge to is not compatible with Acpype. Check the log file and sqm.out under the [filename]_charge_tmp folder. If there is an atom that is not supported, please assign a charge with MOPAC or QM, or enter it manually.
13.12.11. Q. The process of assigning the force field takes a long time and never ends. Is there a solution?
Selecting GAFF or OPLS-AA/L+GAFF can take a long time because Winmostar uses acpype for parameter assignment and acpype can be slow. For macromolecules with nearly 10,000 atoms per molecule, we have confirmed that the process can still complete normally in about 5–6 hours. Runtime depends on your environment and may be faster or slower. We plan to speed up force-field assignment for these types in future versions.
13.12.12. Q. I get an error message when assigning a force field. Is there a solution?
It is very likely that the molecule you are trying to assign force field to is not compatible with Acpype. Check the log file and sqm.out under the [filename]_charge_tmp folder. If there is an atom that is not supported, try a versatile force field such as UFF or Dreiding.
13.12.13. Q. How do I assign a point charge to a polymer?
A. Polymers have large molecular weights, and using the point charge assignment method commonly used for small molecules would require quantum chemical calculations for huge molecules, which would require extremely long calculation times and would be impractical. Also, polymers assume various conformations in vacuum, and the effect of these conformations would be significant when performing quantum chemical calculations for the entire molecule, but the model in the classical force field for polymers does not want extreme charge changes between repetition units. For this reason, polymers are usually assigned point charges in a different way than small molecules. Winmostar’s Polymer Builder allows you to assign point charges in the state of the repeating unit, which are then used to automatically configure the polymer charge. Details can be found in Homo Polymer builder , Block Polymer Builder , Random Polymer Builder.
13.12.14. Q. Can I calculate dielectric constant from Gromacs or LAMMPS (Molecular Dynamics Calculation)?
A. Since the dielectric constant is a physical property dependent on the frequency of the external field, and the mechanism is different for each frequency band, we cannot answer them all.
The dielectric constant calculated from Winmostar’s Gromacs and LAMMPS is a component derived from molecular orientation on the assumption that intramolecular polarization does not change with time.
And among them, it is the infinitely slow low frequency limit value calculated from the fluctuation of the dipole moment of the whole system within the simulation time of molecular dynamics calculation.
Be careful for substances with large molecular weights and slow relaxation, such as polymers, because only information within the range that can be observed within the simulation time is known.
The dielectric function calculated from Winmostar’s Quantum ESPRESSO is the dielectric function of the high frequency component derived from the polarization of the electron with the atomic coordinates fixed.
It is necessary to plan the calculation considering how to obtain the experimental value of the dielectric constant to be compared, the properties of the material, and the research purpose.
In addition, it is possible to help with planning with our paid support.
13.12.15. Q. When I try to run an MD calculation from a structure created with the interface builder, the force field assignment fails or the MD calculation fails. Is there a solution?
A. If the distance between the layers is too short, the molecules of the two layers will collide, and unwanted covalent bonds will be created based on the interatomic distance, or a large force will be applied at the start of the MD calculation, causing the MD calculation to fail. The best way is to increase the value of [Direction]-[Interval] in the interface builder and take a long time to equilibrate
13.12.16. Q. When pressure control (constant NPT or NPH) is used, the calculation fails in the middle of the process. Is there a solution?
A. First of all, in a situation where the interaction between molecules is extremely weak, such as in the gas phase or when other phases are dispersed in the gas phase, the pressure control is not stable, so try a method without pressure control. Next, look at the change in density over time in the broken calculation to see what’s going on. Also, the energy, temperature, and pressure must be sufficiently equilibrated at a constant density before pressure control can be performed. At the end of equilibration at constant density, the pressure (the mean value of) is preferably 0 or negative. If the pressure value is large at the end of equilibration at a constant density, the system size will change drastically immediately after starting pressure control. If you want to reduce the pressure at the end of equilibration at a constant density, reduce the initial density (Approximate 50% of final density, possibly larger or smaller). If this does not solve the problem, (1) change the pressure control to Berendsen method instead of Parrinello-Rahman (Nose-Hoover) method, (2) increase the time constant of the pressure control, and (3) divide the calculation with the pressure control into several short calculations.
13.12.17. Q. How do I decide interaction parameters between liquid (organic) and solid (inorganic) in LAMMPS and Gromacs?
A. First, collect as much information as possible what experimental facts are known and related research about the liquid-solid interface. In particular, it is desirable to know the atomic structure of the surface (e.g., Miller’s index, functional groups, etc.) and the chemical properties of the surface (rough things such as wettability, hydrophobic or hydrophilic).
If the LJ parameter and charge values are listed in the appropriate relevant research literature, it is best to use those values. In Winmostar, Set values as follows.
- For the charges, save them once in mol2 format and rewrite the values of the charges to the values in the paper.
- For LJ parameters, use the [Assign Force Field] function to enter the values for each element from [Exception].
If the appropriate values were not listed in the literature, but some chemical properties are known, run simulations with several different charge and LJ parameters, compare the results of the calculations and adopt the parameters that seem reasonable.
If only the atomic structure of the surface is known, the parameters are determined with ab initio calculations (Quantum ESPRESSO).
In some cases, the Lowdin charge function of Quantum ESPRESSO can be used to determine the charge. However, when using lowdin charges, the total value of the charges does not equal zero, so adjustment of the values (e.g., shifting charges of all atoms) is necessary. It is also important to note that the polarization may be overestimated in Lowdin charges.
For LJ parameters, use known parameters (e.g., Dreiding, UFF, CLAYFF, etc.) or calculate them from ab initio calculations and algorithms such as Force Matching.
This complexity of the liquid (organic) - solid (inorganic) interaction parameters stems from the following facts.
1. The interaction parameters (epsilon, sigma, and charge) depend on the atomic position in classical MD, however, The approximation error is larger in the organic-inorganic interface system than in the organic-organic system.
2. Although the surface of inorganic materials in real devices is covered with oxide films, it is not easy to observe them experimentally and it is difficult to model accurately at atomic resolution.
13.12.18. Q. When I run Gromacs, I get the error “There is no domain decomposition for 49 ranks that is compatible with the given box and a minimum cell size of…”. error. How should I deal with this?
A. Most MD simulation software divides the simulation cell during parallel computation and assigns the divided region (subregion) to each process. Gromacs (and other major software) is designed in such a way that each edge of this subregion cannot fall below the cutoff radius of the uncoupled interaction (the near component of LJ or Coulomb). This error means that each edge of the subregion is below the cutoff radius. There are several ways to deal with this problem, but basically, the number of molecules (number of atoms) should be increased or the number of parallelism should be reduced to a level where the error does not occur. Although there may be requests to increase the number of parallels to the limit, in reality, if the small area becomes smaller than a certain level, the amount of communication increases in relation to the amount of calculation, and the parallel efficiency drops significantly.
13.12.19. Q. When I load the final structure or animation from Gromacs, or when I edit it in some way and then run the MD calculation again, the force field assignment fails or the MD calculation breaks down. Is there a solution?
A. There is a possibility that the molecules are separated due to a bug in Gromacs. To determine if a molecule is separated or not, go to [Select]-[Select by molecular species] and see if there are any unexpected components. If it is separated, please refer to Molecules that should be connected to each other are displayed separately in the structure or the final structure after loading the Gromacs trajectory..
13.12.20. Q. When loading the final structure or animation in Gromacs, the molecules are sometimes displayed in pieces.
13.12.21. Q. Even when Gromacs ER method result loading is executed, the result cannot be displayed / an error will appear.
A. Please check the contents of ermod.out generated in the output destination directory specified when executing the ER method.
If it says “The minimum of the energy coordinate is too large” in ermod.out, press the [Options] button on the ER method execution window and select [For Solution System] Please reduce the value of [minimum value of the solute-solvent energy (ecdmin)].
For details on how to set concrete values, please see ERmod wiki FAQ.
Similarly, check the contents of ermod.out and ERmod wiki FAQ and if you need to change the setting of ermod, set it in [Options] of ER method execution window.
13.12.22. Q. The error message “Out of range atoms - cannot compute PPPM” appears in the LAMMPS log and the calculation terminates abnormally. Please let me know how to deal with this.
A. That message is displayed when an MD calculation fails in a general sense. See Q. I get an error while running a molecular dynamics calculation, what should I do?.
13.12.23. Q. How can I plot the trajectory of a specific atom?
A. Animation operation area Use Options→Tools→Draw Path. Alternatively, follow the steps below.
1) Display animation
2) Click the atom for which you want to draw a trajectory in the molecular display area.
3) Click [Custom Plot] in the animation operating window.
4) Check the second item under [X axis] and select “Atomic position (x)” (choose an appropriate direction as appropriate).
5) Click [Apply] under [X axis].
6) Check the second item under [Y axis] and select “Atomic position (y)” (choose an appropriate direction as appropriate).
7) Click [Add] under [Y axis].
8) Uncheck [Y axis] except for “Atomic position (y)”.
9) Uncheck [Autoscale] from [X Axis] and [Y Axis] below the graph and set [Min] and [Max] accordingly.
In addition, we are planning to develop a function that simplifies the above operations.
13.12.24. Q. The dissipative particle dynamics (DPD) calculation gave me a different result than intuitive. Why is this?
There is no single standard procedure for approximating real molecular assemblies with a DPD model, and the issue is often that the chosen model is not adequate for the target system. DPD interactions are typically coarse and may be too simplified for some systems. Coarse-grained model construction (including DPD) is highly case-dependent, and a generic “quick fix” often does not solve the underlying problem. In Winmostar support, we typically address such cases through individual consulting, starting from requirements and model design.
13.13. About Quantum ESPRESSO, OpenMX
13.13.1. Q. I installed the Quantum ESPRESSO pseudopotential file as per the instructions, but it is not recognized.
A. Copy the pseudopotential files with the .UPF extension (e.g.
O.pw-mt_fhi.UPF, Au.pbe-dn-rrkjus_psl.0.1.UPF, etc.) to the pseudo folder under the Quantum ESPRESSO installation folder ( By default, this is C:\Program Files\Quantum ESPRESSO 64-bit 5.2.1pseudo etc.) under the Quantum ESPRESSO installation folder. However, if you download and save the UPF file using a browser such as the old Internet Exproler, there have been reports of problems such as the extension being changed without your permission, so please also try a browser such as Edge or Chrome.13.13.2. Q. I don’t know how to find a pseudopotential file for Quantum ESPRESSO. How do I find them?
A. It is described in Quantum ESPRESSO for Windows Installation Manual.
13.13.3. Q. Calculation using Quantum ESPRESSO fails. An error appears in displaying the calculation result.
A. First of all, please take action of general errors.
Next, since Winmostar continuously executes each module of QE by batch processing, we look at the flow of processing described in Winmostar generated bat file (local execution) or sh file (remote execution) While checking the generated output file (pwout or out) file in order.
For example, check ph.x output log (ph.out) for phonon calculation.
Please take action on the place where error such as “Error in routine …” appears first, then execute the job again.
Errors related to specific keywords, please confirm the setting of that keyword official site <https://www.quantum-espresso.org/Doc/INPUT_PW.html>.
Typical QE error handling is described in official site FAQ <http://www.quantum-espresso.org/resources/faq>.
If you change to nosym=.True. and the calculation flows, you may be experiencing a bug that originates from the detection of symmetry in the atomic configuration.
Try reducing the number of parallels; the calculation may then proceed normally. Changing the diagonalization method (david, cg, rmm-*, etc.) may also help. Because this behavior depends on how Quantum ESPRESSO was built (compiler, MPI, numerical libraries), it can be useful—especially for remote jobs—to rebuild Quantum ESPRESSO with different settings if needed. You may also improve stability by specifying the parallel distribution options (-nk, -nt, -nb, -nd) in the Options for executable section of the Options tab in the Quantum ESPRESSO Keyword Settings window.
13.13.4. Q. When executing Phonon calculation using Quantum ESPRESSO, “third order derivatives not implemented with GGA” is output in the ph.x output (ph.out), and the calculation result cannot be obtained.
A. This can be solved by selecting a pseudopotential other than GGA.
13.13.5. Q. Quantum ESPRESSO, OpenMX SCF calculation or structure optimization calculation does not converge.
A. Please carry out the following countermeasures in order.
Always to try:
· First-principles calculations have many setting items, so do not change the calculation conditions appropriately, and stream a series of calculations while recording properly.
Take action for general problem of QE.
· Plot the Estimated accuracy against the number of SCF cycles for QE. It is still good if it is a log-log plot.
· Spin Polarization State · Determine whether electric charge is appropriate.
· Gives the order of up/down spins.
· If you shift K_POINTS in Hexagonal crystal, remove the shift.
· Constrain the magnetic moment of the whole system.
· Use a reasonable initial structure.
· Use structures obtained by experiments and other calculation methods.
· Check to see if the calculation converges with a simpler model, such as a surface model only for an interface (adsorption) model, a bulk model for a surface model, or a bulk model for a defect model.
· If there are atoms whose placement is arbitrary in the calculation (light elements invisible in X-rays, solid solutions, defects, non-integer compositions, slab surface end structures, etc.), try a different placement.
· For slabs, reconsider the cutout and modification of the termination structure.
· In cases where solid solution · defects are included, make the initial structure so that large dipole moment does not occur in the system.
Try next:
· Adjust mixing_mode and diagonalization.
· Make mixing_beta smaller.
· Increase nbnd.
· Change the type of pseudopotential file.
· Adjust the initial value of spin polarization. (Atomic unit or whole system)
· If the entire system may have a dipole moment, such as in a slab model, consider using a dipole correction or using the ESM method (bc=1) or changing the surface composition.
· If the U parameter is used, gradually increase it from U=0 to check the behavior.
· If relatively complicated conditions such as external electric field, defect, adsorption, etc. are set, try them under simpler conditions than eliminating them, and if the calculation converges, the final state of the calculation (atomic arrangement · wave function Etc.) as the initial state and starts the calculation.
· Calculation is started from halfway of computation that did not converge (SCF algorithm depends on history).
Try in balance with calculation time / calculation precision:
· Take a large cutoff energy.
· Take a lot of K points.
Try with balance with calculation accuracy:
· For spin-polarized calculations, first check convergence without spin polarization.
· Adjust smearing (use/not use, type, width).
Loosen the convergence parameter of the SCF.
13.13.6. Q. I cannot visualize the three-dimensional distribution of electron density, spin density, and potential in Quantum ESPRESSO.
A. First trace the Quantum ESPRESSO Basics section of the Winmostar tutorial and see if you can display them on a small system size. If not, please refer to Q. I cannot create a model as I expected. It cannot calculate. It does not work. and rebuild your environment. If the tutorial works fine but does not, first check if a cube file has been generated in your working folder, if so, this may be due to a problem with Winmostar Viewer. If a cube file is generated, please refer to Cannot open large size cube file in Winmostar Viewer. to fix the problem.
13.13.7. Q. The SCF calculation of Quantum ESPRESSO is displayed as “too few bands” in the output file (.pwout or .out) and abnormally ends. I do not know how to set nbnd.
A. First of all, please check nbnd description of the QE official manual.
If you do not use nbnd when you do the calculation, QE automatically calculates nbnd by setting it automatically, so please uncheck “Use nbnd” on the keyword setting screen of Winmostar.
To increase nbnd, set nbnd to a value larger than the value of “number of Kohn-Sham states” that is output to the pwout or out file when executed without using nbnd.
Also please refer to the value of “# valence bands:” displayed at “Use nbnd” on Configure window of Winmostar (See Solid ‣ Quantum ESPRESSO menu for detail).
13.13.8. Q. The SCF calculation of Quantum ESPRESSO terminates abnormally with “fixed occupations and lsda need tot_magnetization” displayed in the output file (.pwout or .out). How do I resolve this?
A. Select smearing in the occupations section, or use tot_magnetization instead of starting_magnetization.
13.13.9. Q. Quantum ESPRESSO’s structural optimization calculation (vc-relax) terminates abnormally with the output file (.pwout or .out) showing “smooth g-vectors missing ! in the output file (.pwout or .out) and terminates abnormally. How can I resolve this?
A. Recalculate using the structure output at the end of the pwout or out file. on the Winmostar GUI, create a new job with the structure optimization animation displayed and the final structure.
13.13.10. Q. The SCF calculation of Quantum ESPRESSO on metal is terminated abnormally with “charge is wrong” displayed in the output file (.pwout or .out). How do I resolve this?
A. First try the Q. Quantum ESPRESSO, OpenMX SCF calculation or structure optimization calculation does not converge. procedure.
If the log shows “WARNING: integrated charge= … expected= …” immediately before “charge is wrong,” your charge and/or spin settings may be inconsistent with the model. Adjust settings such as tot_charge, nspin, starting_magnetization, and/or tot_magnetization accordingly. For example, if the total number of electrons in the simulation cell is odd (as can happen in defect or surface models), you might set nspin=2 and tot_magnetization=1.
If this does not solve the problem, make sure that occupations is set to smearing in the Quantum ESPRESSO Keyword Settings window, and if necessary, set ecutrho to a larger value (such as 400 Ry) and the number of k-point divisions to a larger value (gradually increasing) Set the k-point divisions as large (gradually increasing) as necessary.
13.13.11. Q. When calculating the dielectric function using Quantum ESPRESSO, “bad band number” is displayed in the epsilon.x output (eps.out) and it is not possible to acquire the dielectric function.
A. It is solved by increasing the number of bands (nbnd) by SCF calculation.
13.13.12. Q. When calculating the dielectric function using Quantum ESPRESSO, “USPP are not implemented” is displayed in the epsilon.x output (eps.out) and it is not possible to acquire the dielectric function.
A. Select a norm preserving pseudopotential in the SCF calculation.
13.13.13. Q. My Phonon calculation using Quantum ESPRESSO failed and I cannot get the results.
A. First, please do not run the Phonon calculation, but use the same pseudopotential file and other calculation conditions to make sure that the SCF calculation finishes successfully. Next, check the ph.out (the output file of the Phonon calculation module ph.x) in the working directory (the folder ending in _qe_data). If it says “The phonon code with US-PP and raman or elop not yet available” there, you can solve it by selecting a norm-preserving pseudopotential. Also refer to Q. I cannot find a norm-preserving, LDA functional pseudopotential file for Quantum ESPRESSO for the element I want to calculate in the pseudopotential files installed by default. How can I find it? when looking for pseudopotential files. Similarly, if you are using a PAW potential, Raman calculations, etc., are not supported, and you can solve this problem by choosing a norm-preserving pseudopotential. If you are experiencing symmetry-related errors, please run [Solids]-[Refine Lattice] before running the Phonon calculation.
13.13.14. Q. I cannot find a norm-preserving, LDA functional pseudopotential file for Quantum ESPRESSO for the element I want to calculate in the pseudopotential files installed by default. How can I find it?
A. This DB <https://pseudopotentials.quantum-espresso.org/legacy_tables/hartwigesen-goedecker-hutter-pp>`_ contains a relatively large number of norm-preserving and LDA-functional pseudopotential files. The pseudopotential file is not limited to this pseudopotential file. Please verify the cutoff energy and the number of k-points used in your calculations, not limited to this pseudopotential file.
13.13.15. Q. When I try to output the Fermi surface, they do not appear.
A. First, make sure that the target substance is a metal. Next, output the density of states, and confirm that the density of states is not 0 on Fermi energy.
13.13.16. Q. Can I calculate the dielectric constant from Quantum ESPRESSO (band calculation)?
13.13.17. Q. How do you set up the type of functionalities in Quantum ESPRESSO?
A. In Quantum ESPRESSO, since a pseudopotential file is created for each function, the function is basically determined when the pseudopotential file is selected. Some functions (e.g., HSE, vdw functions) select the pseudopotential file created by the base function (e.g., PBE in the case of HSE) and then use the input_dft keyword to override the function’s settings.
13.13.18. Q. Why does the cell size structure optimization in Quantum ESPRESSO result in discontinuous changes or energy increases in the final structure energy?
A. The reason for the slight discontinuity in the energy change at the end is that Quantum ESPRESSO is reworking the plane wave set to use only the last calculation to avoid problems caused by the Pulay pressure.
This energy jump can be reduced to some extent by increasing the cutoff energy.
If the final structure differs significantly from the initial structure, continue and run the structural optimization calculation again.
13.13.19. Q. Quantum ESPRESSO terminates normally for systems with a small number of atoms, but for large systems it terminates abnormally with an error such as “cannot allocate …”. for large systems. What can I do about this?
A. Please refer to the FAQ on memory usage in the official QE documentation.
13.13.20. Q. Why do the values for Total DOS appearing in Projected DOS (PDOS) and DOS appearing in Density of States (DOS) differ in Quantum ESPRESSO?
A. In Quantum ESPRESSO, PDOS is calculated by projecting the electronic states obtained in the plane-wave basis onto the atomic basis, but this projection is not perfect. This is a property of plane wave basis DFT codes in general, not just Quantum ESPRESSO.
13.13.21. Q. When I set Hubbard U in Quantum ESPRESSO for structural optimization, it terminates abnormally with the error message “NR-step length unreasonably short”. What can I do about this?
A. If the version of Quantum ESPRESSO is 5 series, the force calculation for Hubbard U is not implemented and this error occurs. Therefore, upgrading to Quantum ESPRESSO version 6 or later will solve this problem.
13.13.22. Q. Why does the number of cycles required for convergence of SCF calculations before and after Quantum ESPRESSO 6.6 change so much?
A. One reason is that the default value of diago_david_ndim was changed from 4 to 2 in Quantum ESPRESSO 6.6 and later. Even after QE6.6, if diago_david_ndim is explicitly set to 4, the behavior is similar to earlier versions.
13.13.23. Q. When I use hybrid functional (e.g. HSE) in Quantum ESPRESSO, the band structure collapses. Is there a way to deal with this?
A. Unfortunately, as of version 7.1, Quantum ESPRESSO itself does not support band structure calculations with hybrid functionals (mixed functionals). It is possible to do so by post-processing using an external module (such as Wannier90), but Winmostar (as of V11.6.4) does not support such post-processing, so users must perform post-processing on Winmostar’s output results themselves. The user must perform the post-processing on the Winmostar output.
13.13.24. Q. When using hybrid functionals, such as HSE, in QuantumESPRESSO, the DOS display becomes distorted, and the bandgap becomes larger than literature values. Is there a solution?
A. Avoid setting the nqx1/2/3 values too low relative to the k-point count. As a guideline, set the lower limit to 70% of the k-point count. Additionally, align the k-point spacing in the DOS plot with the k-point spacing value in pw.x. Implementing these measures will yield the desired DOS.
13.13.25. Q. Unnecessary bonds are displayed while modeling the system or when reading the calculation results. Or the bonds are not displayed. How can I fix this?
A. See this FAQ for more information.
13.13.26. Q. When setting up workflow and keywords in Quantum ESPRESSO, even if the cell is not converted to a primitive cell, it is converted to a primitive cell when the calculation is executed.
A. Do not check “Use Bravais-lattice index” or “Set ibrav = N and celldm”.
13.13.27. Q. Is there any scaling factor proposed for first-principles calculations (e.g. Quantum ESPRESSO) such that the vibrational spectra (IR, Raman) obtained from calculations such as quantum chemical calculations (e.g. GAMESS, Gaussian) are matched to experimental values?
A. It seems to be not proposed. The reason is that quantum chemical calculations (Gaussian basis) have only Hamiltonian (including functional) and basis functions as parameters, whereas ab initio calculations (plane-wave basis pseudopotential method) have many parameters, and it is difficult to propose a generic scaling factor that takes them into account. There is no guarantee that this will work, but if you are in a situation where you can compare experimental and calculated values, you may want to construct your own scaling factor for a particular combination of pseudopotential, k-point density, cutoff energy, and functional.
13.13.28. Q. When I run a calculation on a local machine with MPI enabled in OpenMX, I get the error tcp_peer_send_blocking: send() to socket 12 failed: Transport endpoint is not connected.
A. For issues specific to Cygwin’s OpenMPI, go to Windows Settings - Network and Internet - Change Adapter Options and disable the network adapter you are not using. Also, OpenMX recommends that you use OpenMP on the local machine.
13.13.29. Q. After running a structural optimisation calculation with a change in cell size in OpenMX, if I run a continuation job with SCF.Restart=On, the standard output shows Failed (3) in reading the restart files, <Restart> Could not find restart files.
A. The error message indicates that the number of charge density meshes in the restart file is different from the number of meshes you are about to calculate. This can happen if the cell size changes significantly during the structural optimisation calculation. In such cases, a new SCF calculation must be performed after the structural optimisation calculation.
13.14. Addons
13.14.1. Q. When calculating the Hansen solubility parameter of a polymer using the solubility parameter calculation module, the output value changes depending on how the repeating structure (monomer) of the polymer is taken.
A. It is because of the group contribution algorithm implemented. When searching for an atomic group, the search is started from the largest atomic group. It is recommended that functional groups that seem to be important be included in the repeating unit.
13.14.2. Q. The value of the Hansen solubility parameter obtained using the solubility parameter calculation module is significantly different from the literature value.
A. The solubility parameter calculation module outputs Hansen solubility parameters using the group contribution method trained by the neural network using various literature values as learning data. Therefore, it does not return exactly the same value as the literature value. Please note that the unit of the solubility parameter differs depending on the literature.