Abstract | ||
---|---|---|
The excessive complexity of both machine architectures and applications have made it difficult for compilers to statically model and predict application behavior. This observation motivates the recent interest in performance tuning using empirical techniques. We present a new embedded scripting language, POET (parameterized optimization for empirical tuning), for parameterizing complex code transformations so that they can be empirically tuned. The POET language aims to significantly improve the generality, flexibility, and efficiency of existing empirical tuning systems. We have used the language to parameterize and to empirically tune three loop optimizations - interchange, blocking, and unrolling - for two linear algebra kernels. We show experimentally that the time required to tune these optimizations using POET, which does not require any program analysis, is significantly shorter than that when using a full compiler-based source-code optimizer which performs sophisticated program analysis and optimizations. |
Year | DOI | Venue |
---|---|---|
2007 | 10.1109/IPDPS.2007.370637 | IPDPS |
Keywords | Field | DocType |
empirical tuning system,parameterized loop optimization,linear algebra kernel,static analysis,authoring languages,program analysis,program diagnostics,compiler-based source-code optimizer,program control structures,optimising compilers,poet embedded scripting language,code transformation,loop optimization,predictive models,scripting language,application software,source code,linear algebra,parameterized complexity,kernel | Kernel (linear algebra),Parameterized complexity,Computer science,Parallel computing,Static analysis,Compiler,Program analysis,Application software,Performance tuning,Scripting language,Distributed computing | Conference |
ISBN | Citations | PageRank |
1-4244-0910-1 | 48 | 1.93 |
References | Authors | |
26 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Qing Yi | 1 | 189 | 13.98 |
Keith Seymour | 2 | 247 | 21.29 |
Haihang You | 3 | 228 | 21.46 |
Richard Vuduc | 4 | 1343 | 100.74 |
Daniel J. Quinlan | 5 | 652 | 80.13 |