srdatalog.compile¶
Public CUDA compile entry point. Thin re-export shim.
The real implementation lives at srdatalog.ir.codegen.cuda.api.
This module exists at the top level so that from srdatalog.compile import ... keeps working for downstream users while the implementation
sits with the CUDA target where it belongs.
Module Contents¶
Data¶
API¶
- srdatalog.compile.__all__¶
[‘Target’, ‘compile_kernel_body’, ‘compile_pipeline’, ‘compile_runner’]