Journal of Parallel and Distributed Computing - Special issue on tools and methods for visualization of parallel systems and computations
ATOM: a system for building customized program analysis tools
PLDI '94 Proceedings of the ACM SIGPLAN 1994 conference on Programming language design and implementation
EEL: machine-independent executable editing
PLDI '95 Proceedings of the ACM SIGPLAN 1995 conference on Programming language design and implementation
Computer aided hand tuning (CAHT): “applying case-based reasoning to performance tuning”
ICS '01 Proceedings of the 15th international conference on Supercomputing
VISTA: a system for interactive code improvement
Proceedings of the joint conference on Languages, compilers and tools for embedded systems: software and compilers for embedded systems
Pinpointing Representative Portions of Large Intel® Itanium® Programs with Dynamic Instrumentation
Proceedings of the 37th annual IEEE/ACM International Symposium on Microarchitecture
Compositional approach applied to loop specialization
Euro-Par'07 Proceedings of the 13th international Euro-Par conference on Parallel Processing
| Hi-index | 0.00 |
Program performance is tightly linked to the assembly code,this is even more emphasized on EPIC architectures. Assessing preciselyquality of compiled code is essential to deliver high performance. Themost important step is to build a comprehensive summary for end-user and extract manageable information. In this paper, we present our firstprototype called MAQAOAdvisor, a key MAQAO(Modular AssemblyQuality Optimizer) module that drives the optimization process throughassembly code analysis and performance evaluation. It performs comprehensiveprofiling, hot-loop and hot-spot detection, fast evaluation andguides local optimizations. An originality of MAQAOAdvisor is to deportpart of optimizations from the driver to a post-compiler evaluationstage. It is based on static analysis and dynamic profile of assembly code.It feeds information back to help end-user detect and understand performanceproblems. It proposes optimization recommendations to guidea user to perform the best transformations to get the best performance.