Loading...
Please wait, while we are loading the content...
Similar Documents
Path-sensitive Value-flow Optimizations
| Content Provider | Semantic Scholar |
|---|---|
| Author | Moir, Mark Schlansker, Mike |
| Copyright Year | 1998 |
| Abstract | Observations of values computed during program execution suggests that compilers miss numerous optimization opportunities: even in highly optimized programs, up to 40% of executed instructions compute the same value as their previous dynamic instance. Very likely, this redundant recomputation reeects expression equivalence that holds only along some|but frequent|execution paths. Because not all executions of an expression are optimizable, this some-paths redundancy is beyond conventional optimizers: conservative analyzers fail to expose it, and innexible transformations fail to remove it. To harness the some-paths potential, this research proposes an optimization framework for exploiting redundancies that may not exist along all execution paths, or may diier along some paths. The underlying framework algorithms will be applicable to a broad class of optimization problems that are based on value ow, i.e. on the paradigm of reusing the ow of previously computed values. As a result, the framework will be capable of deriving path-sensitive versions of common sub-expression elimination, load/store elimination, constant propagation, and other value-ow optimizations. This research will address both analysis and transformation: their power will be obtained by novel integration of techniques with orthogonal strengths. While the goal is to develop algorithms that are complete, they must also permit natural, proole-directed reduction of power and control of compile-time cost. Finally, phrasing value-ow optimizations within a single uniform framework will enable reuse of optimization modules that will be derived from parameterizable algorithms. |
| File Format | PDF HTM / HTML |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
