Dear Bocheng,
I hope this email finds you well. Following your recommendation, I
downloaded the newer PyChrono .conda package and installed it locally via
Conda. However, when I ran a test script that imports pychrono.vehicle, I
encountered an ImportError related to the* _vehicle DLL module*.
The specific error traceback is as follows:
C:\Users\dkq\anaconda3\envs\chrono_test\python.exe
C:\Users\dkq\PycharmProjects\PythonProject2\setupTest-Tutorial-Magic2025\crm\tutorial_VEH_CRMTerrain_WheeledVehicle.py
Traceback (most recent call last):
File
"C:\Users\dkq\PycharmProjects\PythonProject2\setupTest-Tutorial-Magic2025\crm\tutorial_VEH_CRMTerrain_WheeledVehicle.py",
line 2, in <module>
import pychrono.vehicle as veh
File
"C:\Users\dkq\anaconda3\envs\chrono_test\Lib\site-packages\pychrono\vehicle.py",
line 13, in <module>
from . import _vehicle
ImportError: *DLL load failed while importing _vehicle*: 找不到指定的模块。
I would greatly appreciate it if you could provide some insights into the
possible causes of this error and the corresponding solutions. For your
reference, here are the key details of my environment and operations:
My Conda environment is named chrono_test, with Python 3.12 (matching the
PyChrono package's py312 tag).
I installed the local .conda package(py312h84d2775_6211) by running conda
install [local package path] after uninstalling the old PyChrono version.
The operating system is Windows 11 (64-bit).
Thank you very much for your time and help!
Best regards,
Ding
在2025年12月22日星期一 UTC+8 14:07:30<[email protected]> 写道：
> Hi Ding,
>
> Thank you for getting back! I would recommend trying later release, such
> as py312h84d2775_6211 , which should come with FSI support in PyChrono.
>
> Best,
>
> Bocheng Zou
>
>
> On 12/22/25 06:04, Kelvin Ding wrote:
>
> Dear Bocheng,
> Thank you so much for your prompt reply!
> I installed PyChrono through Conda, and the specific version information
> on my side is as follows:pychrono 9.0.1 py312h84d2775_6161 (projectchrono)
> If you have any additional guidance, tips, or relevant documentation
> regarding the usage of FSI in PyChrono, I would be truly grateful for your
> sharing.
> Best regards,Ding
>
> 在2025年12月21日星期日 UTC+8 01:03:09<[email protected]> 写道：
>
>> Hi Ding,
>>
>> Thank you for reaching out! FSI is supported in PyChrono. Could you
>> please help me confirm how did you install PyChrono? Are you installing it
>> through Conda or compiling from scratch?
>>
>> Best,
>>
>> Bocheng Zou
>>
>>
>> On 12/19/25 07:37, beihu dong wrote:
>>
>> Hello everyone!
>> I’m currently working on the simulation and calculation of lunar rover
>> terramechanics. So far, I’ve successfully configured Project Chrono for C++
>> and PyChrono for Python through Windows.
>> However, I’m not as familiar with C++ as I am with Python, so I’m asking
>> everyone: are the current features of PyChrono comprehensive enough to
>> implement the calculation process for FEM-based wheels and SPH-based soil?
>> If yes, what are the installation commands for the latest version of
>> PyChrono? Since I’ve already installed PyChrono 9.0.1, it seems that it
>> doesn’t support the pyChrono::FSI module yet.
>>
>> All the best！
>> Ding
>>
>> [image: 屏幕截图 2025-12-19 152954.png][image: 屏幕截图 2025-12-19 153029.png]
>>
>> 在2025年9月28日星期日 UTC+8 23:50:23<Radu Serban> 写道：
>>
>>> That’s fantastic, Rebeca.
>>>
>>>
>>>
>>> We’ll make a note about the number of threads issue (in case other users
>>> run into similar issues). The truth is we have not used older GPUs in such
>>> a long time, we take for granted certain capabilities.
>>>
>>>
>>>
>>> If you can get access to a cluster with newer GPUs, you will see a
>>> dramatic increase in computational performance.
>>>
>>>
>>>
>>> Best,
>>> Radu
>>>
>>>
>>>
>>> *From:* [email protected] <[email protected]> *On
>>> Behalf Of *Rebeca Guimarães
>>> *Sent:* Saturday, September 27, 2025 4:34 PM
>>> *To:* ProjectChrono <[email protected]>
>>> *Subject:* Re: [chrono] FSI won't work
>>>
>>>
>>>
>>> Hey there, Radu!
>>>
>>>
>>>
>>> I'm thrilled to tell you I was able to run the FSI module in my
>>> computer!
>>>
>>> It turned out FsiForceWCSPH was requesting 1024 threads and my GPU was
>>> unable to give that. I then adapted it and changed to fewer threads and it
>>> worked out.
>>>
>>>
>>>
>>> I'm still working on the visualization, since I only had Irrlicht, so
>>> I'll now install VSG.
>>>
>>>
>>>
>>> It's just a temporary solution, since I'm requesting approval to use my
>>> uni's laboratories' computers, but up until then, I'm able to continue my
>>> VIPER single wheel simulations. Once I have access to the labs, I hope to
>>> be able to run complete rover simulations.
>>>
>>> Em sexta-feira, 26 de setembro de 2025 às 15:28:46 UTC-3, Radu Serban
>>> escreveu:
>>>
>>> Hi Rebeca,
>>>
>>>
>>>
>>> So, you do have a pretty old GPU. Anything before 7.5 is now considered
>>> legacy. In fact, 7.5 is also marked for obsoletion.
>>>
>>> I’m afraid you’ll need to get access to newer hardware.
>>>
>>>
>>>
>>> --Radu
>>>
>>>
>>>
>>>
>>>
>>> *From:* Rebeca Guimarães <[email protected]>
>>> *Sent:* Friday, September 26, 2025 1:48 PM
>>> *To:* Radu Serban <[email protected]>
>>> *Subject:* Mensagem particular referente a: [chrono] FSI won't work
>>>
>>>
>>>
>>> Hey there, Radu!
>>>
>>>
>>>
>>> My last try had indeed been with CUDA 12.3 yet again, since my last try
>>> with 13.0 resulted on the exe simply closing right from the start.
>>>
>>>
>>>
>>> Searching for CUDA
>>>
>>> CUDA compiler: D:/Program Files/NVIDIA GPU Computing
>>> Toolkit/CUDA/v12.3/bin/nvcc.exe
>>>
>>> CUDA toolkit version: 12.3.52
>>>
>>> CUDA toolkit root dir: D:/Program Files/NVIDIA GPU Computing
>>> Toolkit/CUDA/v12.3
>>>
>>> CUDA archs (from cache): 60
>>>
>>> CUDA found and enabled.
>>>
>>> Searching for Thrust
>>>
>>> Thrust version: 2.2.0
>>>
>>> Thrust include dir: D:/Program Files/NVIDIA GPU Computing
>>> Toolkit/CUDA/v12.3/include
>>>
>>>
>>>
>>> I was able to surpass the atomicAdd error by defining sm_60 instead of
>>> all_major as the arch, since I have a GPU with compute capability of 6.1.
>>> On the other hand, now I'm getting a new error of too many resources
>>> requested for launch when I compile and run the single wheel FSI demo. But
>>> that only happens when I run my own compiled demo, the pre compiled demo
>>> available on the Chrono website runs normally.
>>>
>>>
>>>
>>> I don't know if there are any differences between the source code of the
>>> pre compiled demo and latest one available on git that would explain the
>>> higher demand for GPU. Or if it's still an error due to CUDA issues.
>>>
>>>
>>>
>>> If you could shed any light on the matter, it would help a lot!
>>>
>>>
>>>
>>> Thanks,
>>>
>>> Rebeca
>>>
>>>
>>>
>>>
>>>
>>> Em sexta-feira, 26 de setembro de 2025 às 02:53:51 UTC-3, Radu Serban
>>> escreveu:
>>>
>>> I would suggest holding off on using the PyChrono conda package that Dan
>>> mentioned. That is not based on the current Chrono code (and was set up as
>>> a temporary solution for some tutorials we recently ran at UW-Madison).
>>> Updated official PyChrono conda packages with Chrono::FSI support will come
>>> soon.
>>>
>>> --Radu
>>>
>>>
>>>
>>> *From:* [email protected] <[email protected]> *On
>>> Behalf Of *Rebeca Guimarães
>>> *Sent:* Thursday, September 25, 2025 11:30 PM
>>> *To:* ProjectChrono <[email protected]>
>>> *Subject:* Re: [chrono] FSI won't work
>>>
>>>
>>>
>>> Hello, everyone!
>>>
>>>
>>>
>>> Thank you for the suggestion, Dan! I will definetely try out PyChrono. I
>>> need to present a proof of concept to my undergrad thesis advisor in the
>>> next weeks so I had decided to go with C++ since I struggled a bit with
>>> Anaconda. But I'll try again :)
>>>
>>>
>>>
>>> Radu, I've been using the latest version, I downloaded it from the git.
>>> I'm able to compile and build Chrono and run Vehicle module projects.
>>>
>>>
>>>
>>> The FSI single wheel demo that comes in the pre compiled pack works just
>>> fine in my computer. But whenever I try to compile a new project, even the
>>> demo cpp itself, it will run and then simply close the prompt before the
>>> visual interface ever starts.
>>>
>>>
>>>
>>> This has been happening with CUDA 12.3, CUDA 12.9, CUDA 13.0.
>>>
>>>
>>>
>>> What baffles me is the fact I'm able to run the pre compiled exe, so it
>>> doesn't seem to be a hardware incompatibility. I really have no idea what's
>>> going on :(
>>>
>>> Have any of you ever faced a similar problem, where the prompt closes
>>> right at the FSI Initialize() line?
>>>
>>>
>>>
>>> Rebeca
>>>
>>>
>>>
>>>
>>>
>>> Em terça-feira, 23 de setembro de 2025 às 04:46:32 UTC-3, Radu Serban
>>> escreveu:
>>>
>>> Hi Rebeca,
>>>
>>>
>>>
>>> Could you please clarify what Chrono version you are using? Older
>>> releases will obviously not work with new CUDA releases.
>>>
>>>
>>>
>>> The current Chrono code in the main branch of the repository should work
>>> fine with any new CUDA 12.* versions (e.g. 12.8 or 12.9), as well as CUDA
>>> 13.0 (although I only tested the latter on Linux). Any of these CUDA
>>> versions will work fine with the latest VS versions.
>>>
>>>
>>>
>>> Please make sure that:
>>>
>>> - you use the latest Chrono code from the development branch
>>> (‘main’).
>>> - Update your copy of Visual Studio to the latest version.
>>> - You can test with CUDA 13.0.1. If that doesn’t work, please let
>>> me know. But if you can install a different version of CUDA, I suggest
>>> you
>>> use 12.9 which is what I currently use on my Windows machine.
>>>
>>>
>>>
>>> --Radu
>>>
>>>
>>>
>>>
>>>
>>> *From:* [email protected] <[email protected]> *On
>>> Behalf Of *Rebeca Guimarães
>>> *Sent:* Thursday, September 18, 2025 10:45 PM
>>> *To:* ProjectChrono <[email protected]>
>>> *Subject:* [chrono] FSI won't work
>>>
>>>
>>>
>>> Good afternoon!
>>>
>>>
>>>
>>> I've donwloaded and compiled Chrono and been able to run many projects.
>>> Since I'm running terramechanics analysis, I am now trying to run FSI SPH.
>>>
>>>
>>>
>>> I was able to code, configure and geenrate, but when I run the exe, it
>>> will halt at the Initialize and close the cmd.
>>>
>>>
>>>
>>> I think I've narrowed the issue down to CUDA. I'm using version 13.0.1.
>>> I read at the website I need to use 12.3, but my MSVC says it expects
>>> version 12.4 or newer.
>>>
>>>
>>>
>>> Can someone help me how to downgrade my MSVC so it will accept 12.3?
>>>
>>>
>>>
>>> Thanks!
>>>
>>>
>>>
>>> D:\Program Files\Microsoft Visual
>>>
>>> Studio\2022\Community\VC\Tools\MSVC\14.44.35207\include\yvals_core.h(902):
>>> error : static assertion failed with "error STL1002: Unexpected compiler
>>> version, expected CUDA 12.4 or newer."
>>>
>>> --
>>>
>>> You received this message because you are subscribed to the Google
>>> Groups "ProjectChrono" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>> To view this discussion visit
>>> https://groups.google.com/d/msgid/projectchrono/b3ff10c6-ac6c-42f7-bcb5-af0b30085b6fn%40googlegroups.com
>>>
>>> <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/b3ff10c6-ac6c-42f7-bcb5-af0b30085b6fn*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!JjYeUOI4xLXD7t_mTxzl_aANBnQnLjdYX9a-waRJUB_I6ou_X9kc_E7vaOf0TUf33x5i9MicivFtLuWIyXw$>
>>> .
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "ProjectChrono" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>>
>>> To view this discussion visit
>>> https://groups.google.com/d/msgid/projectchrono/96f01d65-23b5-487b-9402-c739d258cd17n%40googlegroups.com
>>>
>>> <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/96f01d65-23b5-487b-9402-c739d258cd17n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!IQJpQJeDKnI0JXIk_rqO2pUeJuza79YSWKwthdIVpRy9_xQ_GY0T1Uu0Q_Kj9poCgCo-Q6lHpH-OAse6K2w$>
>>> .
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "ProjectChrono" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>>
>>> To view this discussion visit
>>> https://groups.google.com/d/msgid/projectchrono/fec67aab-1bfe-452c-a2f1-80ad05f8390dn%40googlegroups.com
>>>
>>> <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/fec67aab-1bfe-452c-a2f1-80ad05f8390dn*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!Jo90yPlde-u3efkp5e6m7QvbgNks5yZYQYlan1cGqjYLWgN1iR4LADTF97_GTMOaDTAO6cPpb6yMo2HE1nc$>
>>> .
>>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "ProjectChrono" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>>
>> To view this discussion visit
>> https://groups.google.com/d/msgid/projectchrono/4d6867b0-2717-4b08-98ab-ec103d1db00bn%40googlegroups.com
>>
>> <https://urldefense.com/v3/__https://groups.google.com/d/msgid/projectchrono/4d6867b0-2717-4b08-98ab-ec103d1db00bn*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!J31WeATp8e4HVmCRsdY4XmgxwnX838z3uQ9j9BwdELSvGDaA_wKNZI3zsAplkKTyLEboq_J01gwgf8ILudbZ$>
>> .
>>
>> --
> You received this message because you are subscribed to the Google Groups
> "ProjectChrono" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
>
> To view this discussion visit
> https://groups.google.com/d/msgid/projectchrono/14e62a8a-e5bc-4f9b-8a04-2387a744df29n%40googlegroups.com
>
> <https://urldefense.com/v3/__https://groups.google.com/d/msgid/projectchrono/14e62a8a-e5bc-4f9b-8a04-2387a744df29n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!J31WeATp8e4HVmCRsdY4XmgxwnX838z3uQ9j9BwdELSvGDaA_wKNZI3zsAplkKTyLEboq_J01gwgf9iBY-_C$>
> .
>
>
--
You received this message because you are subscribed to the Google Groups
"ProjectChrono" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/projectchrono/ef7c5747-182c-45e4-b2a8-75e741f7a215n%40googlegroups.com.
