Vyas Ramasubramani
Vyas Ramasubramani
### Description It would be nice to change all uses of `enum` in freud to `enum class`. ### Motivation and Context Scoped enums are generally preferred to C-style enums since...
## Description The current logic for finding the allocator .so file presumes that torch.py is in the same directory tree as the compiled allocator. While this is true when the...
**Is your feature request related to a problem? Please describe.** Currently librmm packages contain vendored CCCL code because the packages are fetched via CPM. This vendoring behavior is potentially problematic...
**Is your feature request related to a problem? Please describe.** After #1324 the C++ APIs of rmm are documented within the Sphinx documentation along with the Python APIs. As a...
**Is your feature request related to a problem? Please describe.** Classes like `fixed_size_memory_resource` define a lock type as a protected member. This exposes an implementation detail of the class that...
**Is your feature request related to a problem? Please describe.** `statistics_resource_adapter` contains two `std::shared_lock` types that could instead be `std::shared_mutex` now that we require C++ 17. **Describe the solution you'd...
**Is your feature request related to a problem? Please describe.** In the process of working through issues in #1324 we identified a number of improvements that we'd like to see...
**Is your feature request related to a problem? Please describe.** Currently when librmm is built, it accepts a CMake option `CUDA_STATIC_RUNTIME` to determine whether to link to cudart statically or...
**Is your feature request related to a problem? Please describe.** `rmm.reinitialize` will clean up any internal references to memory resources prior to recreating new instances. However, it currently has no...
**Is your feature request related to a problem? Please describe.** Currently rmm intermixes C++ API bindings and Cython declarations in the same pxd files. All of these pxd files are...