John Camier

I understand the issue, and how the parser instantiates the nested kernels later in the code. But what bothers me is that the provided kernel should be in 'his space'...

hum, I was more thinking of http://en.cppreference.com/w/c/preprocessor/conditional

Well, even if they are in the OKL source code, if I do something like: ` #ifdef __NVCC__ {CUDA debug stuff} #else {other host debug} #endif ` then the parser...

Thank you for taking the time to investigate memory utilization. Most objects are allocated and computed during setup and are especially designed to remain on the device to reduce memory...

I looked at the main objects that might be released, and the mesh's cached GeometricFactors could be destroyed or generated in a temporary memory area after the setup stage. There...

The [amr branch](https://github.com/ceed/Laghos/tree/amr) of Laghos does work with partial assembly. I will try it again with these changes.

Tested successfully on the [amr](https://github.com/CEED/Remhos/tree/amr) branch of Remhos: | master | coarsen-L2-proj | |---|---| |  |  | |Mass loss u: 6.10436e-05 | Mass loss u: 2.33147e-14| `mpirun -n...

I am still having issues with `FormSystemMatrix`, the `DiagonalPolicy` and now `EliminateVDofs`. `DiagonalPolicy::DIAG_ONE` seems to be only supported by `FABilinearFormExtension::EliminateBC`: only the comment allows to know this and is not...

Will be re-opened when needed.

Should be ready to test and use.