Hartmut Kaiser

@biddisco FWIW, we do have a latch (see here: https://github.com/STEllAR-GROUP/hpx/blob/master/hpx/lcos/local/latch.hpp). Also, if you don't need the future returned by `hpx::async`, just use `hpx::apply` (i.e. fire and forget) - that saves...

> My question is about how we could nicely fire off tasks and use a latch to keep track of them, without the user having to do anything to their...

I agree. This new executor could even be a wrapper around any other (already existing) executor, just adding the ability to wait for all executed tasks. Also, some of our...

What would that entail? What does it mean 'supporting Windows HPC Pack'?

> > What would that entail? > > probably just need to add another environment class into `libs/core/batch_environments` While I can blindly try adding such a binding, I don't have...

FWIW, the environment variables are described here: https://docs.microsoft.com/en-us/previous-versions/windows/it-pro/windows-hpc-server-2008r2/gg286970(v=ws.10)

hello_world_distributed always works. To test the core bindings, use the `--hpx:print-bind` command line option.

Here is an explanation of how to manually run HPX applications on more than one locality (outside of batch environments): https://stackoverflow.com/a/35381710/269943. I have not tried using ipv6 addresses on the...

I'm not aware of distributed_hello_world being broken, even more as it relies on built-in facilities to establish the connection between the different localities. How can I reproduce your issues? The...

What I might be able to do to debug your problem is to explicitly set the environment variables HPC Pack normally sets. Can you give me those for both localities?