Mapping performance data for high-level and data views of parallel program performance
ICS '96 Proceedings of the 10th international conference on Supercomputing
Performance measurement tools for high-level parallel programming languages
Performance measurement tools for high-level parallel programming languages
Efficient management of parallelism in object-oriented numerical software libraries
Modern software tools for scientific computing
Proceedings of the 14th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications
Semicoarsening Multigrid on Distributed Memory Machines
SIAM Journal on Scientific Computing
A sound type system for secure flow analysis
Journal of Computer Security
The role of instrumentation and mapping in performance measurement
The role of instrumentation and mapping in performance measurement
LLVM: A Compilation Framework for Lifelong Program Analysis & Transformation
Proceedings of the international symposium on Code generation and optimization: feedback-directed and runtime optimization
Data centric cache measurement using hardware and software instrumentation
Data centric cache measurement using hardware and software instrumentation
An API for Runtime Code Patching
International Journal of High Performance Computing Applications
The Tau Parallel Performance System
International Journal of High Performance Computing Applications
Dynamic inference of abstract types
Proceedings of the 2006 international symposium on Software testing and analysis
Quantitative information flow as network flow capacity
Proceedings of the 2008 ACM SIGPLAN conference on Programming language design and implementation
Assigning Blame: Mapping Performance to High Level Parallel Programming Abstractions
Euro-Par '09 Proceedings of the 15th International Euro-Par Conference on Parallel Processing
Hi-index | 0.00 |
Traditional methods of performance analysis offer a code centric view, presenting performance data in terms of blocks of contiguous code (statement, basic block, loop, function). Data centric techniques, combined with hardware counter information, allow various program properties including cache misses and cycle count to be mapped directly to variables. We introduce mechanisms for efficiently collecting data centric performance numbers independent of hardware support. We create extended data centric mappings, which we call variable blame, that relates data centric information to high level data structures. Finally, we show performance data gathered from three parallel programs using our technique.