|
Uni20
Asynchronous tensor and runtime library.
|
Files | |
| conjugate_inplace.hpp | |
| CUDA reference backend for in-place scalar conjugation. | |
| copy.hpp | |
| CUDA runtime backend for contiguous Tensor transfers. | |
| copy_task.hpp | |
| Coroutine kernel-dispatch customization for CUDA Tensor copies. | |
| elementwise_arithmetic.hpp | |
| Compiled CUDA entry points for named stateless arithmetic transforms. | |
| elementwise_conjugate.hpp | |
| Typed launch boundary for CUDA accessor-respecting in-place conjugation. | |
| elementwise_copy.hpp | |
| Typed launch boundary for CUDA accessor-respecting element copies. | |
| elementwise_plan.hpp | |
| CUDA lowering for compact affine elementwise iteration plans. | |
| elementwise_scale.hpp | |
| Compiled CUDA entry points for elementwise scaling. | |
| mdspec_traits.hpp | |
| Shared CUDA reference-backend traits for raw buffer mdspecs. | |
| partitioned_buffer_linear.hpp | |
| Allocation-wide CUDA linear operations for partitioned buffers. | |
| transform.hpp | |
| CUDA reference backend for registered elementwise transforms. | |