Hello ODM Team,
I installed ODM natively on my ubuntu 16.04 server. After installation I tried to test the odm with the auckerman dataset. It started well but stopped after meshing with MVS Textturing Cell:
[INFO] Running ODM Meshing Cell - Finished
[INFO] Running MVS Texturing Cell
[DEBUG] Writing MVS Textured file in: /home/derfred/ODMProjects/TEST001/odm_texturing/odm_textured_model.obj
[DEBUG] running /home/derfred/OpenDroneMap/SuperBuild/install/bin/texrecon /home/derfred/ODMProjects/TEST001/opensfm/reconstruction.nvm /home/derfred/ODMProjects/TEST001/odm_meshing/odm_mesh.ply /home/derfred/ODMProjects/TEST001/odm_texturing/odm_textured_model -d gmi -o gauss_clamping -t none
/home/derfred/OpenDroneMap/SuperBuild/install/bin/texrecon (built on Jun 27 2018, 08:31:21)
Load and prepare mesh:
PLY Loader: comment PCL generated
Reading PLY: 100104 verts… 199788 faces… done.
Generating texture views:
NVM: Loading file…
NVM: Number of views: 76
NVM: Number of features: 0
Loading 0%…Illegal instruction (core dumped)
Traceback (most recent call last):
File “OpenDroneMap/run.py”, line 47, in
plasm.execute(niter=1)
File “/home/derfred/OpenDroneMap/scripts/mvstex.py”, line 139, in process
‘{keepUnseenFaces}’.format(**kwargs))
File “/home/derfred/OpenDroneMap/opendm/system.py”, line 34, in run
raise Exception(“Child returned {}”.format(retcode))
Exception: Child returned 132
So it exited with error 132. I checked the community pages and found out that this comes up when the hardware might be to old.
I’m running it on a HP proLiant DL385 G6 rack server with a six core AMD Opteron Processor Model 2435 (2.6 GHz) CPU and 62GB ram.
So is there any experience to handle this problem?
@pierotofy I tried using the rebuild docker instead of the natively installation. It exited on the same point during meshing again with the 132 error as posted above.
Could you open an issue on https://github.com/OpenDroneMap/OpenDroneMap/issues so that we can track this and add paste a link to this thread? If possible, it would also be nice to have a copy of the dataset you are using (if you can share / upload it on google drive or dropbox).
Just clutching at a straws, but have you tried using the -march GCC flag for your chipset (https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html)? A few options on setting flags are here (https://stackoverflow.com/questions/10085945/set-cflags-and-cxxflags-options-using-cmake) - the suggestion for using an env variable maybe the easiest if it works, as you should then just be able to run the script (rather than stopping the install and editing CMakeFile). Haven’t played at this detailed level for compiler options in years (not sure how it will go) but one one the other pro’s here can probably help if necessary.