Kevin Williams
Trinity College, Dublin
7 Papers
51 Citations
Kevin Williams is an academic researcher from Trinity College, Dublin. The author has contributed to research in topics: Compiler & SIMD. The author has an hindex of 6, co-authored 7 publications. Previous affiliations of Kevin Williams include French Institute for Research in Computer Science and Automation.
Chat about Author
Papers
Vapor SIMD: Auto-vectorize once, run everywhere
Dorit Nuzman,Sergei Dyshel,Erven Rohou,Ira Rosen,Kevin Williams,David Yuste,Albert Cohen,Ayal Zaks +7 more
- 02 Apr 2011
TL;DR: This work presents a synergistic auto-vectorizing compilation scheme that leverages the optimized intermediate results provided by the first stage across disparate SIMD architectures from different vendors, having distinct characteristics ranging from different vector sizes, memory alignment and access constraints, to special computational idioms.
Dynamic interpretation for dynamic scripting languages
Kevin Williams,Jason McCandless,David Gregg +2 more
- 24 Apr 2010
TL;DR: This paper presents a novel intermediate representation for scripting languages that explicitly encodes types of variables in a flow graph, where each node is a specialized virtual instruction and each edge directs program flow based on control and type changes in the program.
An experimental study of sorting and branch prediction
TL;DR: This paper empirically examining the behavior of the branches in all the most common sorting algorithms finds insertion sort to have the fewest branch mispredictions of any comparison-based sorting algorithm, and that bubble and shaker sort operate in a fashion that makes their branches highly unpredictable.
Vectorization technology to improve interpreter performance
Erven Rohou,Kevin Williams,David Yuste +2 more
- 20 Jan 2013
TL;DR: This work introduces a novel approach for interpreter optimization that reduces instruction dispatch thanks to vectorization technology and extends the split compilation paradigm to interpreters, thus guaranteeing that this approach exhibits almost no overhead at runtime.
15
Speculatively vectorized bytecode
Erven Rohou,Sergei Dyshel,Dorit Nuzman,Ira Rosen,Kevin Williams,Albert Cohen,Ayal Zaks +6 more
- 24 Jan 2011
TL;DR: The proposed Vapor SIMD first applies complex ahead-of-time techniques to vectorize source code and produce bytecode of a standard portable format, yielding up to 14.7x and 11.8x speedups on x86 and PowerPC platforms (including JIT-compilation time).