duneuro-bst issueshttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues2020-10-15T16:00:37Zhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/17test for memory leak?2020-10-15T16:00:37ZJuan GPCtest for memory leak?Hi, as mentioned in a previous thread,
I've been overriding new operator to understand the use of the memory api both in duneuro-bst and in duneuro in general.
So far the new new operator isn't covering all cases and thus this is just ...Hi, as mentioned in a previous thread,
I've been overriding new operator to understand the use of the memory api both in duneuro-bst and in duneuro in general.
So far the new new operator isn't covering all cases and thus this is just some ideas.
1. I've seen a bid difference between memory allocated and memory deleted so far. I need to work more on this.
2. Because of 1, ive seen how the memory of the whole OS behaves when having subsequent hundreds of calls to duneuro-bst and the accessible memory does in fact get reduced every call. However this can happen for other reasons.
3. There are hundreds of memory allocations and deletes in the program. Perhaps a more unified access to memory would help to actually make the application run faster.https://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/16cell not found in neighborhood of seed2020-09-24T07:33:26ZTakfarinas Medanicell not found in neighborhood of seedHello all, @juan.garciaprieto, @christi
We are testing some other scenario with duneuro,
In this case, it was for a sEEG setup where the electrodes and the sources are within the wm and grey matter area,
I have tried with the Venant a...Hello all, @juan.garciaprieto, @christi
We are testing some other scenario with duneuro,
In this case, it was for a sEEG setup where the electrodes and the sources are within the wm and grey matter area,
I have tried with the Venant and the PI source model, and the error is the same.
(using the CG with the std options)
Here is the observed error, is it familiar to you? do you have have any idea how to fixed it?
```
....
dipole_3970.time = 0
dipole_3971.density = sparse
dipole_3971.time = 0
dipole_3972.density = sparse
dipole_3972.time = 0
dipole_3973.density = sparse
dipole_3973.time = 0
dipole_3974.density = sparse
dipole_3974.time = 0
Searching -5.05585e-005 0.0495549 0.0106144, seed at 0.000206744 0.0496615 0.0108159 (maxDist: 0.000687487)
Dune reported error: Dune::Exception [searchNeighborhood:/mnt/c/Users/33649/Documents/MATLAB/duneuroWindows/bst-duneuro/src/duneuro/duneuro/common/kdtree.hh:238]: cell not found in neighborhood of seed (1 0)
```
Thanks in advanceChristian EngwerChristian Engwerhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/15Coarse solver did not converge (MEG)2020-09-24T08:27:38ZTakfarinas MedaniCoarse solver did not converge (MEG)Hi, @christi @juan.garciaprieto @sophie.schrader
We have just noticed a new error, this one appears in the case of the high tetra mesh density,
for the MEG, (with the CG, venant model + std parameters).
we have never reached this st...Hi, @christi @juan.garciaprieto @sophie.schrader
We have just noticed a new error, this one appears in the case of the high tetra mesh density,
for the MEG, (with the CG, venant model + std parameters).
we have never reached this step for this model since the MEG fails before,
since we have solved the high memory cach problem, now the program achieve this step and then fails,
here is the error
```
DUNEURO> Error log:
Reading 3d Gmsh grid...
version 2.2 Gmsh file detected
file contains 786143 nodes
file contains 4334286 elements
number of real vertices = 786143
number of boundary elements = 0
number of elements = 4334286
volume_conductor.time_reading_gmsh = 322.639
tensors = 5
time = 0.000997
volume_conductor.time_reading_tensors = 0.000997
volume_conductor.time_reordering_indices = 0.152637
volume_conductor.time = 322.793
degree = 1
element_type = tetrahedron
WARNING: You are using AMG without SuperLU! Please consider installing SuperLU, or set the usesuperlu flag to false to suppress this warning.
time = 0.0219416
time = 0.0219428
time = 0.709101
This is a new model. Compute the transfer Matrix.
solver.coil_0.projection_0.time_rhs_assembly = 5.39961
=== matrix setup (max) 5.21023 s
solver.coil_0.projection_0.linear_system_solver.time_matrix_setup = 5.21023
solver.coil_0.projection_0.linear_system_solver.time_matrix_assembly = 15.631
solver.coil_0.projection_0.linear_system_solver.iterations = 404
solver.coil_0.projection_0.linear_system_solver.reduction = 9.88265e-011
solver.coil_0.projection_0.linear_system_solver.conv_rate = 0.944571
solver.coil_0.projection_0.linear_system_solver.time_solution = 196.472
solver.coil_0.projection_0.linear_system_solver.time = 217.314
solver.coil_0.projection_0.linear_system_solver.time = 217.366
solver.coil_0.projection_0.time_solution = 217.366
solver.coil_0.projection_0.time = 222.765
solver.coil_1.projection_0.time_rhs_assembly = 4.03322
solver.coil_1.projection_0.linear_system_solver.iterations = 99
solver.coil_1.projection_0.linear_system_solver.reduction = 9.40078e-011
solver.coil_1.projection_0.linear_system_solver.conv_rate = 0.791988
...
...
solver.coil_151.projection_0.linear_system_solver.iterations = 41
solver.coil_151.projection_0.linear_system_solver.reduction = 4.81438e-011
solver.coil_151.projection_0.linear_system_solver.conv_rate = 0.560215
solver.coil_151.projection_0.linear_system_solver.time_solution = 20.1182
solver.coil_151.projection_0.linear_system_solver.time = 20.1192
solver.coil_151.projection_0.linear_system_solver.time = 20.158
solver.coil_151.projection_0.time_solution = 20.158
solver.coil_151.projection_0.time = 24.273
solver.coil_152.projection_0.time_rhs_assembly = 4.12796
**Dune reported error: MathError [mgc:/home/juan/bst-duneuro/src/dune-istl/dune/istl/paamg/amg.hh:837]: Coarse solver did not converge**
***************************************************************************
** Error: Error during the DUNEuro computation, see logs in the command window.
***************************************************************************
```
Have you seen this error before, any ideas about the reason?
Thanks in advance.Christian EngwerChristian Engwerhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/14WARNING : AMG without SuperLU!2020-10-15T16:14:24ZTakfarinas MedaniWARNING : AMG without SuperLU!Hi,
just a small point, I have noticed this message in the actual version but I have never had the chance to go deeper,
`WARNING: You are using AMG without SuperLU! Please consider installing SuperLU, or set the usesuperlu flag to fal...Hi,
just a small point, I have noticed this message in the actual version but I have never had the chance to go deeper,
`WARNING: You are using AMG without SuperLU! Please consider installing SuperLU, or set the usesuperlu flag to false to suppress this warning.`
is it relevant to discuss it or to add it?/Takfarinas MedaniTakfarinas Medanihttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/13document binary file format2020-09-17T21:28:01ZChristian Engwerdocument binary file formatWe need a proper documentation of the binary file format. The reader has potential for improvement, but the format is not documented.
In particular it should be possible to simply mmap most of the content of the file.We need a proper documentation of the binary file format. The reader has potential for improvement, but the format is not documented.
In particular it should be possible to simply mmap most of the content of the file.Juan GPCJuan GPChttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/12check use of drivers and precomputed info2020-09-17T04:43:46ZJuan GPCcheck use of drivers and precomputed infohttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/11solve dangling pointer issue2020-09-17T04:39:18ZJuan GPCsolve dangling pointer issuehttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/10copy and generate a minimal version based on what we have working already2020-09-17T04:39:07ZJuan GPCcopy and generate a minimal version based on what we have working alreadyhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/9integrate and improve interruption management2020-09-17T04:38:46ZJuan GPCintegrate and improve interruption managementhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/8rearange computations so that they use less memory. Old memory<>time tradeoff.2020-09-17T04:38:20ZJuan GPCrearange computations so that they use less memory. Old memory<>time tradeoff.https://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/6FEATURE: read/write conductivity files with binary format2020-09-17T04:36:00ZJuan GPCFEATURE: read/write conductivity files with binary formathttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/5define optimal list of flags for the compiler2020-09-17T04:35:35ZJuan GPCdefine optimal list of flags for the compilerhttps://gitlab.dune-project.org/duneuro/brainstorm/duneuro-bst/-/issues/3Dipole localization failes for high density meshes2020-09-24T02:10:26ZChristian EngwerDipole localization failes for high density meshesIn the case of EEG, with high tetra density mesh, the main error is this one:
```
...
dipole_38805.time = 0
dipole_38806.density = sparse
dipole_38806.time = 0
dipole_38807.density = sparse
dipole_38807.time = 0
Dune reporte...In the case of EEG, with high tetra density mesh, the main error is this one:
```
...
dipole_38805.time = 0
dipole_38806.density = sparse
dipole_38806.time = 0
dipole_38807.density = sparse
dipole_38807.time = 0
Dune reported error: Dune::Exception [findEntityImpl:/home/juan/bst-duneuro/src/duneuro/duneuro/common/edgehopping.hh:84]: coordinate is outside of the grid, or grid is not convex
```
It seems that is related to the dipoles located outside the grid, but I have checked the location of these dipoles,
and I ensure that they are all within the brain compartment (GM)
Also, could you please explain quickly what this error means: grid is not convex?
The source space is defined just by an independent set of dipole(s),
is there any test to do on the source space (these dipoles) to avoid this error?