Denis Demidov

Serialization is not currently possible in amgcl. I am not even sure it is possible to implement it in a generic way: amgcl moves preconditioner data (matrices and vectors) to...

> I guess the way to do it would be to serialize whatever internal structures are built right before sending them to the backend. Yes, that is a viable idea....

Yes, there is a wrapper code using pybind11 in the [pyamgcl](https://github.com/ddemidov/amgcl/tree/master/pyamgcl) folder. Also see these examples: - [examples/pysolver.py](https://github.com/ddemidov/amgcl/blob/master/examples/pysolver.py) - [examples/pyprecond.py](https://github.com/ddemidov/amgcl/blob/master/examples/pyprecond.py)

That is a good point, but I am currently away from keyboard for about two weeks, and so won't be able to implement it right away. I'll gladly accept a...

Something like this? https://vexcl.readthedocs.io/en/latest/interop.html

Yes, vex::vector may be initialized with `vex::device_vector`: https://vexcl.readthedocs.io/en/latest/memory.html#_CPPv4N3vex6vector6vectorERKN7backend13command_queueERKN7backend13device_vectorI1TEE6size_t and `vex::device_vector` may be constructed from `cl::Buffer`: https://github.com/ddemidov/vexcl/blob/5097dc6b6614235dcf2ce59f402e63618dfbaab1/vexcl/backend/opencl/device_vector.hpp#L66

The following works for me: ```cpp #include int main() { vex::Context ctx(vex::Filter::Env && vex::Filter::Count(1)); std::cout

The test you mentioned is defined *either* for OpenCL 2.0 or the CUDA backend. I don't think NVIDIA supports OpenCL 2.0.

Hi Rene, The main design goal of VexCL was to make GPGPU developement as painless as possible, but since VexCL is based on standard API(s), and does not hide the...

One way I can think of making this possible, is to introduce a thin wrapper for the expression being assigned to. Say, something like `vex::let(lhs_expr, event_wait_list = null, event =...