René Widera

> Edit: in the code, the comment of the class called from inside this macro looks outdated, we will fix. The macro should work as described in the docs. will...

This is not a problem because the pointer is not used to free memory. We use alpaka buffers which handle the lifetime and locations of the buffer correctly.

As @sbastrakov said setting`mallocMC_ALPAKA_PROVIDER` should work. In picongpu we still use a very old [cmake find module for mallocMC](https://github.com/ComputationalRadiationPhysics/cmake-modules/blob/dev/FindmallocMC.cmake) In general, it should be possible to use mallocMC without this...

Compile tests can be performed in the current github runners. But to test runtime CUDA and HIP must be executed in our HZDR gitlab CI.

I suggest using the apt repositories. If you like to build ROCm you should use spack because AMD is maintaining spack better than their documentation on how to build ROCm....

There is no time to work on this issue :-(

@sabreshao Is there a time frame when this feature will be supported? We need this feature for our visualization library [ISAAC](https://github.com/ComputationalRadiationPhysics/isaac/tree/dev) too.

> @psychocoderHPC hcc is deprecated and rocm is switched to hip-clang. > @scchan Do we support function pointer array in hip-clang? The error with HIP clang is still the same....

The reason for the long kernel names @kloppstock got is that he is encoding many information into the name, thats the same we do with PIConGPU. Not sure how @SimeonEhrig...

> During the accessor development #1249 I needed to implement a few meta functions on the side. Since alpaka is TMP heavy, we are going to need such metaprogramming facilities...