Michael Schellenberger Costa
Michael Schellenberger Costa
Looks like we are running into nvbug4824645 This should be fixed in a future release
That is unfortunate, the test was passing locally :(
closing in favor of #4534
@Aminsed sorry about the delay, this is just a matter of prioritization
From my point of view we should first consider where cuCollections should live. Given the general plan to unify CCCL more I would suggest into moving it libcu++. That is...
> * please note that `std::pair/cuda::std::pair` is not even trivially copyable. [@miscco](https://github.com/miscco) maybe we can think about a `cuda::pair` version focused on performance. `cuda::std::pair` is trivially copyable alignment is an...
/ok to test cb9806f
> I thought we'd already taken care of this here: #1249 > > Or does this PR do more than that? This is `trivially_relocatable` which is a non standard extension...
in #2127 I am making tuple `trivially_copyable`when possible The same is already true for pair, so do we actually need this at all? Or does thrust not default to `is_trivially_copyable`
> GCC 6 rightfully complains about: > > ``` > In file included from /home/coder/cccl/libcudacxx/lib/cmake/libcudacxx/../../../include/cuda/std/complex:25:0, > from /home/coder/cccl/thrust/testing/type_traits.cu:12, > from /home/coder/cccl/build/cuda11.1-gcc6/thrust-cpp14/thrust/testing/thrust.cpp.omp.cpp14/type_traits.cu.cpp:1: > /home/coder/cccl/libcudacxx/lib/cmake/libcudacxx/../../../include/cuda/std/detail/libcxx/include/complex:1470:96: error: literal operator suffixes not preceded by...