Mikael Simberg

I tried to follow the discussion here but I may have missed the answer somewhere. Do I understand correctly that the statuses mentioned in the description > ExecutionSpaceStatus::submitted > ExecutionSpaceStatus::running...

> @msimberg Yes, the statuses don't need to be exposed to the user (anymore). The question for you is if there is any functionality in `HPX` that lets you query...

It seems like HIP 5.7 triggers the error: ``` /opt/spack-6cfcbc01679ebabe640f4462c8fc07626a543015/opt/spack/linux-ubuntu22.04-zen3/gcc-12.2.0/stdexec-git.nvhpc-23.09.rc4_main-fu7ialflbf3rvcwd2zxjs5fg64c72mlf/include/stdexec/execution.hpp:6339:44: error: out-of-line definition of 'operator()' does not match any declaration in 'get_completion_scheduler_t' auto get_completion_scheduler_t::operator()(const _Queryable& __queryable) const noexcept ^~~~~~~~ ```...

I'll return to this once the CSCS maintenance is over and I can test more. For anyone bumping this into the meantime, using pika with stdexec and HIP requires using...

I'll use this branch to try bumping the spack commit used with HIP 5.5 first (which is failing on other PRs), then get back to bumping HIP from 5.5 to...

Curiously things start failing with https://github.com/spack/spack/commit/d0589285f72bc6ec9b5f804612448954dcdf819c. Makes no sense to me at the moment...

https://github.com/spack/spack/pull/45788 should work around the build issues. It seems like there's some sort of issue caused by a new version of ninja, though I don't know what exactly triggers it....

stdexec simply doesn't compile with HIP 5.7: https://compiler-explorer.com/z/jEYz9qjTv. Since 5.6 and older, as well as 6.0 and newer don't fail the same way I'm going to assume this is simply...

Hey, thanks for having a look at this. I'll admit I don't know yet if this will necessarily be a useful thing to implement, but I opened the issue to...

#1042 enabled valgrind testing for all modules. The only ones not tested are the MPI and CUDA/HIP modules.