stdgpu
stdgpu copied to clipboard
stotko
stdgpu: Efficient STL-like Data Structures on the GPU
In contrast to boost, thrust and others, stdgpu is not a header-only library and, hence, requires shipping a compiled library. The following module currently require source file compilation: - ~~`bitset`:...
Our `bitset` class is a GPU version of `std::bitset` which, however, is designed to cover more use cases. In particular, its interface and implementation (run-time fixed-sized) is somewhere between `std::bitset`...
Until now, the header names follow a simple convention. There are essentially three different types of extensions: - `.h`: Header files that can be used in both CUDA (`.cu`) and...
Don't know if this feature already exists but do unordered_maps support structures other than like . Similarly can unordered_set support using a hashing function from boost::hash or vectors using your...
**Is your feature request related to a problem? Please describe.** I'd like to use this in a hardware vendor agnostic way, and more specifically I'd like to use with webgpu....
The backend system is currently restricted to build and install the library only for a single backend. **Current behavior**: - Set `STDGPU_BACKEND` to either `STDGPU_BACKEND_CUDA` (default) or `STDGPU_BACKEND_OPENMP` to control...
Up to now, the container classes have a fixed capacity and are created using the non-standard `createDeviceObject` factory function. Furthermore, since ease of use in GPU kernels is considered a...
**Describe the bug** unordered map creation freezes async processes **Steps to reproduce** ```c++ runBuldKernel > > (ng, object_size_ui); // The line below would only complete when runBuldKernel is done stdgpu::unordered_map...
### Describe the bug Building VS project failed when the backend is CUDA 12.5. ### Steps to reproduce 1. Prerequisites: 1. Windows 11, 2. CUDA 12.5, 3. MSVC 19.41 (VS...
I notices that some functions like `stdgpu::detail::memcpy` is non-async and running on DEFAULT cuda stream. More details: stdgpu::detail::memcpy depends on `dispatch_memcpy` and it looks like: ```c++ dispatch_memcpy(void* destination, const void*...
