Martin Lücke
Google, University of Edinburgh
Actions
Martin is a PhD student at the University of Edinburgh working on compiler construction with a focus on declarative compiler-specific languages. He has contributed to MLIR and several derived projects, namely Rise and xDSL. He is currently an intern at Google DeepMind working on MLIR.
Driving MLIR compilation from Python
The MLIR infrastructure supports productive IR construction via Python bindings, but offers only limited string parsing-based pass pipeline constructor to transform the IR from Python. We propose a Python-native interface to compose and fine-tune transformations at finer granularity by leveraging MLIR’s transform dialect. We also extend this dialect to allow for constructing “passes” that apply a composition of rewrite patterns on-the-fly. This talk highlights the challenges MLIR’s radically extensible design presents for the conventional pass-based design of compiler pipelines and offers a glimpse of the possible evolution.
Martin Lücke
Google, University of Edinburgh
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top