Incremental compilation pipelines process code chunk-by-chunk by building an ever-growing translation unit. Code is then lowered into the LLVM IR and subsequently run by the LLVM JIT. Such a pipeline allows creation of efficient interpreters. The interpreter enables interactive exploration and makes the C++ language more user friendly. The incremental compilation mode is used by the interactive C++ interpreter, Cling, initially developed to enable interactive high-energy physics analysis in a C++ environment.
Our group puts efforts to incorporate and possibly redesign parts of Cling in Clang mainline through a new tool, clang-repl. The project aims to generalize the IncrementalCUDADeviceCompiler of Cling and add this functionality in clang-repl.
There are several forseen tasks:
- Write a detailed request for comment (RFC) document on the design choices and gather feedback from the LLVM community.
- Implement the necessary functionality to support existing test cases available here.
- Develop clang-repl-based tutorials for the CUDA backend.
- Investigate the requirements for supporting a HIP backend.
- Present the work at the relevant meetings and conferences.
- Necessary knowledge: C++ programming; data structures and algorithms.
- Basic Knowledge of Clang and LLVM
- Experience with GPU programming would be an asset.
- Demonstrate necessary functionality for existing test cases.
- Demonstrate a CUDA-executed gradient computed by the Clad automatic differentiation plugin.
Interested students please contact the mentors for an evaluation task.
- Difficulty level (low / medium / high): high
- Duration: 350 hours
- Mentor availability: May-September