Define 2d model class and Simmetrix GeomSim model loader #96
Conversation
cwsmith
changed the title
simmetrix does not store loops, only loop uses, remove them for now
compiles and executes
no leaks or errors under valgrind
cwsmith
changed the title
|
/runtests |
|
Test Result: failure (details) |
|
/runtests |
|
Test Result: failure (details) |
|
/runtests |
|
Test Result: failure (details) |
|
/runtests |
|
Test Result: failure (details) |
OMEGA_H_CHECK is a no-op when NDEBUG is defined
|
/runtests |
|
Test Result: success (details) |
| @@ -165,6 +166,7 @@ if(Omega_h_USE_SimModSuite) | |||
| set(CMAKE_MODULE_PATH ${OLD_CMAKE_MODULE_PATH}) | |||
| set(Omega_h_SOURCES ${Omega_h_SOURCES} Omega_h_meshsim.cpp) | |||
| set(Omega_h_SOURCES ${Omega_h_SOURCES} Omega_h_matchMeshsim.cpp) | |||
| set(Omega_h_SOURCES ${Omega_h_SOURCES} Omega_h_simModel2d.cpp) | |||
There was a problem hiding this comment.
We should probably consider using list(APPEND Omega_h_sources ...) (https://cmake.org/cmake/help/latest/command/list.html#append) before we end up with a bunch of calls to set. There are other places that can benefit from this too
There was a problem hiding this comment.
I agree. Thank you. Draft PR #97 addresses this.
src/Omega_h_mesh2d.cpp
Outdated
| Model2D Mesh2D::updateModel() { | ||
| model = Model2D::MeshModel2D_load(*this); | ||
| return *model; |
There was a problem hiding this comment.
How does the dereferencing operator work here? It is my understanding that Model2D::MeshModel2D_load returns a Model2D not a pointer to it.
There was a problem hiding this comment.
Good catch. Thank you. Fixed with 86e2b4a.
|
@cwsmith is this ready to look at? I can take a look over the weekend. |
|
@jacobmerson Yes. This is all set. Thank you. |
| std::fill(array.data(), array.data()+array.size(), init); | ||
| return array; | ||
| } | ||
|
|
There was a problem hiding this comment.
these constructors can also be used, cant they ?
Line 164 in ef317f8
There was a problem hiding this comment.
Lines 162 to 166 in ef317f8
For
HostWrite(...), the closest ctor I see is the one that takes an initializer list, but I don't think that will support setting all the values in the array when its length is set at runtime. I also thought there would have been one that initialized all the values but I didn't find it.Interestingly, there are
Write() and Read() ctors that support it:Line 64 in ef317f8
Line 95 in ef317f8
There was a problem hiding this comment.
@cwsmith how about if you use stride=0 and set initial value as the offset ?
There was a problem hiding this comment.
for this one
HostWrite(LO size_in, T offset, T stride, std::string const& name = "");
There was a problem hiding this comment.
Yeah, good call. I misunderstood what that ctor did. Thanks.
For the record, it creates a HostWrite from a call to the Write(...) ctor with the same signature.
Lines 223 to 226 in ef317f8
That in turn calls
fill_linear(...).Lines 67 to 73 in ef317f8
I'll create a PR that adds a docstring for that API.
| OMEGA_H_CHECK(model.faceToLoopUse == expected_f2lu); | ||
| } | ||
|
|
||
| int main(int argc, char** argv) { |
There was a problem hiding this comment.
so iiuc, at the moment this functionality added by this PR loads the smd file and queries and stores connectivity and coordinate information in Omega_h::Model2D.
There was a problem hiding this comment.
That's pretty close. It reads the model topology and some of the geometry from the smd file into the Omega_h::Model2D struct. The arrays in the Model2D struct are currently only accessed for a correctness check for a specific test model.
The goal is to query the Model2D data during adaptation to improve the geometric model approximation. That development is going to take some time and I didn't want to have this code stuck in a branch too long.
|
looks good to me overall, some comments attached |
|
notes from discussion with @jacobmerson
|
Adds a Model2D class and support for filling the Model2D data structures from a Simmetrix GeomSim model (.smd).
will_fail_test_func(...)CMake macro)