-
Notifications
You must be signed in to change notification settings - Fork 17
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
issue with dumpWalls2XDMF #65
Comments
Hello, could you paste here the full stack trace? this should be dumped in the log files |
Stack trace (most recent call last): |
It looks like you did not provide the required number of ranks; this setup is 1x1x1 subdomains and needs postprocess rank so you need to launch a 2 ranks simulation (2 ranks per gpu). |
How do I specify the number of ranks. Is it with the -np flag for mpirun? when I run
it runs with 1 rank. I have one gpu so I'd like to run 2 ranks |
you should refer to your mpi documentation; but I would expect that |
I was able to run it with two ranks, but it still crashes at the same line, and know my log files look like (log_00000.log): 08:05:28:932 Rank 0000 at /home/patrick/code/ymero/src/core/logger.h:249 Compiled with maximum debug level 10 and (log_00001.log): 23:20:49:591 Rank 0001 at /home/patrick/code/ymero/src/core/logger.h:249 Compiled with maximum debug level 10 |
Hey @shafferpr! We'd need just a little more information from you.
Since you can run without the wall dump, it may possibly be an issue with the linking to hdf5 library, or incompatible version. I think we can figure this out pretty quick. Dmitry |
Hi @dimaleks, thanks for your help, I changed the debug level and ran the command: The end of the log_00000.log file is here (sorry I can't post the whole thing here, it's an 11 MB file): The output I get directly from YMeRo is here: |
I pushed a possible fix, can you try with the latest commit on master? |
Thanks Lucas, I suspect I'm linking the wrong hdf5 as well but I can't figure out how to link it properly. In ccmake, I had to configure the variable "HDF5_C_LIBRARY_hdf5" by hand.
When I configure with these variables, I get the following warning from cmake:
This seems relevant to me but I can't trace this error message very far. Is it possible that I need to confirm that my python install also has parallel hdf5 libraries installed? |
Hey guys, I think I have a solution here. I was using a conda distribution of python, which does not come with parallel hdf5 by default, so I had to install it with:
Once I installed that, all the cmake defaults worked and I didn't have to change any of those configuration variables. Thanks for your help! |
Great! |
I'm having a problem with the walls.py example from the tutorials: https://ymero.readthedocs.io/en/latest/user/tutorials.html#walls-py
When I attempt to run this script it crashes at the line:
u.dumpWalls2XDMF([wall], h = (0.5, 0.5, 0.5), filename = 'h5/wall')
with the error:
#31 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/pythonrun.c", line 978, in PyRun_FileExFlags [0x5651b23d2090] #30 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/pythonrun.c", line 1025, in run_mod [0x5651b23d1c93] #29 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/ceval.c", line 731, in PyEval_EvalCode [0x5651b23303fb] #28 | Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/ceval.c", line 4187, in PyEval_EvalCodeEx Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/ceval.c", line 4166, in _PyEval_EvalCodeWithName [0x5651b232f668] #27 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/ceval.c", line 3351, in _PyEval_EvalFrameDefault [0x5651b2380078] #26 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Python/ceval.c", line 4861, in call_function [0x5651b235c74d] #25 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Objects/methodobject.c", line 294, in _PyCFunction_FastCallKeywords [0x5651b230502f] #24 Source "/home/conda/feedstock_root/build_artifacts/python_1551342612670/work/Objects/methodobject.c", line 231, in _PyCFunction_FastCallDict [0x5651b22cc693] #23 Source "/home/patrick/code/ymero/src/extern/pybind11/include/pybind11/pybind11.h", line 627, in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) [0x7f2efa98cfa6]\n 624: // 6. Call the function. 625: try { 626: loader_life_support guard{}; > 627: result = func.impl(call); 628: } catch (reference_cast_error &) { 629: result = PYBIND11_TRY_NEXT_OVERLOAD; 630: }
I can run the script without the above line, but I'd really like to export the xdmf file for visualization purposes. Any idea what the problem is? I'm running on Ubuntu 18.04, with python 3.6, gcc 6.5, and cuda 10.1
The text was updated successfully, but these errors were encountered: