Steven Novack, Joseph Hummel, and Alexandru Nicolau. A simple mechanism for improving the accuracy and efficiency of instruction-level disambiguation. In Proc. of the 8th Intl. Workshop on Languages and Compilers for Parallel Computing (LCPC), pages 289--303, Columbus, OH, August 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....structures. Indeed, pointer based data structures are traditionally notorious for resulting in poor performance mainly due to high data cache miss ratios resulting from a fragmented allocation of these data structures in memory. Approaches developed at UC Irvine, including semantic retention [34] and sophisticated pointer analysis [35] allow for intelligent management of dynamic and pointer based data structures to radically improve data cache hit ratios (e.g. through data layout) Operating System Support for Memory Management. With increasing amounts of on chip multiprocessing and ....

S. Novack, J. Hummel, and A. Nicolau. A Simple Mechanism for Improving the Accuracy and Efficiency of Instruction-Level Disambiguation. In Languages and Compilers for Parallel Computing (August 1995.

....registers, for example: global bits32 hp ebx ; bits32 hplim esi ; We remarked in Section 2 that the front end may know a great deal about (lack of) aliasing between memory access operations. We do not yet have a way to express such knowledge in C , but an adaptation of Novack, Hummel, and Nicolau (1995) looks promising. 4 The problem of run time support When a front end and back end are written together, as part of a single compiler, they can cooperate intimately to support high level run time services, such as garbage collection, exception handling, profiling, concurrency, and debugging. In ....

Novack, Steven, Joseph Hummel, and Alexandru Nicolau. 1995. A Simple Mechanism for Improving the Accuracy and Efficiency of Instruction-level Disambiguation, chapter 19. Lecture Notes in Computer Science. Springer Verlag.

....integration in the PROMIS compiler itself. A structural overview of the PROMIS Prototype is shown in Figure 3. The prototype uses Parafrase 2[11] as a front end for doing parsing, symbolic and dependence analysis[4] and source level parallelization, and uses the EVE Mutation Scheduling compiler[6, 9, 5, 7, 10, 8] as a back end for performing instruction level parallelization and code generation. Because both Parafrase 2 and EVE are capable of generating and using an IR based on a HTG, they are well suited to be used for this research on this subject. The effectiveness of the HTG representation can be seen ....

....assertions for use by EVE, which continues the parallelization process at the instruction level for the VLIW targets. Clearly this prototype, with its independent front end and back end, does not realize our ultimate goal of fully integrating the two. However, the Semantics Retention mechanism [7] in EVE does provide partial integration by allowing higher level semantics information derived by Parafrase 2 to be retained for use by EVE at the instruction level. Furthermore, EVE is able to exploit source level transformations performed by Parafrase 2 that increase the availability of ....

Steve Novack, Joseph Hummel, and Alexandru Nicolau. A simple mechanism for improving the accuracy and efficiency of instruction-level disambiguation. In Proceedings of the 8th Annual Workshop on Languages and Compilers for Parallel Computing, Lecture Notes in Computer Science No. 1033, pages 289--303. SpringerVerlag, August 1995.

....integration in the PROMIS compiler itself. A structural overview of the PROMIS Prototype is shown in Figure 6. The prototype uses Parafrase 2[17] as a front end for doing parsing, symbolic and dependence analysis[8] and source level parallelization, and uses the EVE Mutation Scheduling compiler[11, 14, 10, 12, 15, 13] as a back end for performing instruction level parallelization and code generation. Parafrase 2 and EVE are uniquely suited for use in the prototype for several reasons: 1. The intermediate representations of both compilers are based on the Hierarchical Task Graph (HTG) 5] which will also form ....

....can be made on the fly in response to changing resource constraints and availability. This context sensitive approach to code generation allows EVE to adapt to different application domains, target architectures and cost vs. performance trade offs. 5. EVE has a mechanism, called Semantics Retention[12], for retaining higher level semantics information in order to improve the accuracy and efficiency of instruction level disambiguation. Semantics Retention encodes important higher level semantics information in the form of a hierarchical decomposition of the program s (symbolic) address space. ....

Steve Novack, Joseph Hummel, and Alexandru Nicolau. A simple mechanism for improving the accuracy and efficiency of instruction-level disambiguation. In Proceedings of the 8th Annual Workshop on Languages and Compilers for Parallel Computing, Lecture Notes in Computer Science No. 1033, pages 289--303. Springer-Verlag, August 1995.

No context found.

Steven Novack, Joseph Hummel, and Alexandru Nicolau. A simple mechanism for improving the accuracy and efficiency of instruction-level disambiguation. In Proc. of the 8th Intl. Workshop on Languages and Compilers for Parallel Computing (LCPC), pages 289--303, Columbus, OH, August 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC