Skip to main content

> ML_LITERATURE // TILLET-2019-TRITON-INTERMEDIATE-LANGUAGE-COMPILER-TILED-NEURAL-NETWORK-COMPUTATIONS_v1.0

Triton: An Intermediate Language and Compiler for Tiled Neural Network Computations

Philippe Tillet, H. T. Kung, David Cox · ACM SIGPLAN International Conference on Programming Language Design and Implementation (PLDI) (2019)

hardware-compiler2019industry-standardnotAssessed
DOI: 10.1145/3314221.3314642Back to Literature

Principal Contribution

Invented the Triton programming language and compiler, allowing developers to write high-performance GPU kernels in Python without CUDA C++.

Operational Relevance

Directly guides deployment choices and architecture selection for task-kernel-optimization, task-model-compilation.

Assumptions

  • Standard empirical regularity and statistical stability hold across evaluation domains

Limitations

  • Performance characteristics depend on domain distribution and compute allocation parameters

Connected Algorithms, Architectures & Tools

Related Algorithms:
Related Architectures:
Implementing Libraries: