cxx

We're seeing a lot of warnings on Linux, host compiler GCC 9.3.0 with -Wconversion.

For a basic example, compile:

#include <thrust/device_vector.h>
int main() {
  thrust::device_vector<int> a;
  return 0;
}

With: nvcc main.cu -Xcompiler=-Wconversion.

This will result in around 1000 lines of output warnings.
Tested with the Thrust versions in CUDA 11.0, 11.3, and the la

Expected Behavior

I can pass std::vector<int32_t>, but the the c++ library I want to use uses int, not int32_t

std::vector<int> build_array() {
    auto values = std::vector<int> {1,2,3,4,5};
    for (auto& v: values) v = v * 2;
    return values;
}

Not sure how to get around this since I can't impl VectorElement for c_int due to the orphan rule? There must b

When compiling example/device/example_device_radix_sort.cu with

clang++-11 -DTHRUST_IGNORE_CUB_VERSION_CHECK -I../.. --cuda-gpu-arch=sm_61 -std=c++17 -x cuda -c example_device_radix_sort.cu

it errors:

clang: warning: Unknown CUDA version 11.0. Assuming the latest supported version 10.1 [-Wunknown-cuda-version]
In file included from example_device_radix_sort.cu:46:
In file inc