Mikael Simberg
Mikael Simberg
Closing in favour of a more specific issue #1124.
It may be possible to dispatch to the default implementation using the new `transform_sender` mechanism in stdexec (https://ericniebler.github.io/wg21/D2999R1.html). The process would look like this: - customize `schedule_from` for `thread_pool_scheduler`/`thread_pool_domain` via...
For `nvc++` this would require this issue to be resolved: https://gitlab.kitware.com/cmake/cmake/-/issues/23003.
> For `nvc++` this would require this issue to be resolved: https://gitlab.kitware.com/cmake/cmake/-/issues/23003. Until the above is resolved I think we could try using `nvc++` to compile everything, as we used...
`nvc++` seems to still not be able to compile a lot of things, even without stdexec, C++20, or even CUDA device code. We may be able to compile the subset...
nvcc from CUDA 12 at least requires some changes to fmt (it's constexpr detection isn't taking into account the fact that CUDA 12 supports C++20, at least partially). I have...
Lowering the priority of this. The remaining part of this task requires support for nvcc in stdexec, and that simply does not work yet. The only thing applications can do...
Thanks @yurivict. Do you know if this is broken temporarily or "permanently" (meaning there is no plan to fix it so that the old code still works)? Do you know...
This can be done as part of cleaning up the `functional` module in #16.
The main blocker here is supporting annotations through the standard library facilities.