Compiling Second-Order Accelerate Programs to First-Order TensorFlow Graphs
Summary
Hardware acceleration is the method of accelerating calculations with hardware specifically designed for the type of calculations. Accelerate and TensorFlow are libraries that make this accessible to many programmers, but these libraries differ in the level of abstraction and targeted hardware. This thesis investigates the possibility of compiling and executing Accelerate programs in TensorFlow. A compiler is introduced that converts second-order Accelerate programs to first-order TensorFlow graphs, covering 68% of the Accelerate language.
Collections
Related items
Showing items related by title, author, creator and subject.
-
Automatic task parallelism in Accelerate
Matser, Marien (2023)Parallel execution has a large potential to improve the speed of a program. To make good use of parallelism we need a program that is properly split into several tasks that can be performed in parallel. High level frameworks ... -
Verified Compiler Optimisations
Dral, Joris (2022)This thesis explores how to formally verify the correctness of a certifying compiler where the correctness of individual compiler runs is specified by translation relations, each of which characterises the admissible ...