|
Uni20
Asynchronous tensor and runtime library.
|
Files | |
| buffer.hpp | |
| Typed CUDA device buffers and scoped stream access guards. | |
| cuda_error.cpp | |
| cuda_error.hpp | |
| Structured CUDA runtime failures and checked-call helpers. | |
| cuda_error_presentation.hpp | |
| Presentation adapter for structured CUDA runtime failures. | |
| device.cpp | |
| device.hpp | |
| CUDA device discovery and immutable hardware capability snapshots. | |
| initialization.hpp | |
| Stream-ordered diagnostic initialization of CUDA scalar storage. | |
| resource_pool.hpp | |
| Fixed-capacity pools and move-only leases for CUDA provider resources. | |
| runtime.cpp | |
| runtime.hpp | |
| CUDA device, stream, resource, and scoped runtime primitives. | |
| task_awaiters.hpp | |
| CUDA-specific awaiters for device and runtime resource scheduling. | |