libcudacxx
libcudacxx copied to clipboard
NVIDIA
[ARCHIVED] The C++ Standard Library for your entire system. See https://github.com/NVIDIA/cccl
Currently structured bindings for `cuda::std::tuple` and `cuda::std::array` were broken. The reason for that is that the standard requires, that the specializations of `tuple_size` and `tuple_element` reside in namespace std. whereas...
I'm running into issues where `cuda::std::tuple` does not seem to support structured binding declarations. Is this a feature that should work but is broken? Is this unsupported for `cuda::std::tuple` specifically?...
This implements the NVIDIA internal proposal for a `any_resource_ref` abstraction that facilitates the handling of memory allocations on host and device side. To facilitate adoption the proposal is backported to...
In addition to porting the span implementation from libc++ the PR does the following: * Remove outdated support for `tuple` interface * Remove outdated support for `const_iterator` * Change `index_type`...
@jrhemstad edit: This was originally from the Thrust repo about `thrust::complex`. I pointed @NickKarpowicz at `cuda::std::complex` as it will eventually replace `thrust::complex`. @NickKarpowicz reported that `cuda::std::complex` was even slower than...
This also ports all the standard concepts from libc++ preserving the new modular organization. This will be a base for further development such as `memory_resource` or in the far future...
Addresses issue #306 pending verification that none of these checks break standards conformance. As noted in the linked issue, a large performance penalty can happen in certain applications that heavily...
The `operator*(cuda::std::complex, cuda::std::complex)` does an [`isnan`](https://github.com/NVIDIA/libcudacxx/blob/aff156af14d94acf4afd5f8b780e8e6ef050d711/include/cuda/std/detail/libcxx/include/complex#L639) check on the resulting real and imaginary parts of the operation before returning the resulting `complex`. Is this required by the C++ Standard? At...
With a bit of hacking something like the below allows tests to pass. I find that I must still manually edit the `lit.site.cfg` to correctly point to `libcuda.so`. Will try...
This PR moves the "memory model" related docs to a separate page, so that they can be easily linked from the Programming Guide. It also updates the memory model docs....
