W. W. Carlson and J. M. Draper. Distributed data access in AC. In Proc. 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP'95, pages 39--47, Santa Barbara, California, July 1995. IDA Supercomputing Research Center.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....fully automatically or to augment programmer directives. The broader EARTH C project has also clearly demonstrated the value of identifying local private data to drive analyses such as redundant read write removal and communication optimization [37] Among the unsafe (C derived) languages, AC [10], PCP [8] and UPC [11] offer shared and private data. However, their type systems do not distinguish the addresses of private data from narrowed global pointers to shared data. In effect, these languages offer only global shared and local mixed. Although private data exists at run time, the ....

W. W. Carlson and J. M. Draper. Distributed data access in AC. In Proc. 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP'95, pages 39--47, Santa Barbara, California, July 1995. IDA Supercomputing Research Center.

....A barrier is also included to provide synchronization for operations on distributed containers. The following sections describe the components of the run time system. 3. 1 Global Pointers and Global References Global pointers are based on the global type in languages like CC [4] AC [3] and Split C [7] A global pointer to an object of type T is defined as a templated class HPCxxGlobalPtr T p; Global pointers can be passed between contexts to allow a processor to read and modify objects on a remote node. Two basic operations are defined on global pointers. returns a ....

William W. Carlson and Jesse M. Draper. Distributed data access in AC. In Fifth ACM Sigplan Symposium on Principles and Practices of Parallel Programming, 1995.

....program annotations to specify architecturespecific implementation details in otherwise portable parallel codes. This approach is closer in spirit to Titanium s type qualifiers, although the two languages use rather di#erent mechanisms to model and achieve distributed computing. Orca [8] AC [9], and Split C [11] share Titanium s notion of local versus remote data, with Orca placing greater emphasis on compiler driven placement while AC and Split C give the programmer more direct control. The creators of AC observed that locality qualifications applied to multidimensional arrays and ....

William W. Carlson and Jesse M. Draper. Distributed data access in AC. In Proc. 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP'95, pages 39--47, Santa Barbara, California, July 1995. IDA Supercomputing Research Center.

....a parallel recursive divide and conquer style as suggested e.g. in [8] as well as data parallelism, task farming, and other parallel algorithmic paradigms. In contrast, most other languages adopting the SPMD model, like Denelcor HEP Fortran [14] EPEX Fortran [12] PCP [4] Split C [11] AC [7], support only one level of parallelism and one global name space; only PCP has a hierarchical group concept similar to that of ForkLight. Moreover, in ForkLight, control synchronous execution can locally be relaxed towards totally asynchronous computation as desired by the programmer, e.g. for ....

W. W. Carlson, J. M. Draper. Distributed Data Access in AC. In Proc. ACM SIGPLAN Symp. on Principles and Practices of Parallel Programming, pages 39--47. ACM Press, 1995.

....pioneering work by Shasha and Snir [20] which was later extended by Midkiff et al. [17] to handle array based accesses. Neither of these included implementations and the algorithms as presented were not practical because synchronization behavior is ignored. Related to our work is the AC compiler [6], which uses the non blocking memory operations on the Cray T3D. However, since the AC compiler does not employ cycle detection, the compiled code could potentially generate executions that are not sequentially consistent. In our analyses, we analyze the synchronization accesses in the program to ....

W. W. Carlson and J. M. Draper. Distributed Data Access in AC. In ACM Symposium on Principles and Practice of Parallel Programming, July 1995.

....parallelism, but repeated spawning of threads at run time is an expensive operation. Under some circumstances, fork join parallel programs can be automatically converted to SPMD programs [15] In contrast, SPMD languages like Denelcor HEP Fortran [23] EPEX Fortran [16] PCP [7] Split C [14] AC [10] start program execution with all available threads and keep their number constant during program execution. The program is often partitioned into parallel and serial sections separated by implicit barriers, i.e. execution is control synchronous at the top level of program control. Typically, ....

W. W. Carlson and J. M. Draper. Distributed Data Access in AC. In Proc. ACM SIGPLAN Symp. on Principles and Practices of Parallel Programming, pages 39--47. ACM Press, 1995.

....where standard optimizations (like code motion and dead code elimination) cannot be directly applied. Analysis for these programs is based on the pioneering work by Shasha and Snir [18] which was later extended by Midkiff et al. [16] to handle array based accesses. We analyze the synchronization[5] accesses in the program and obtain precedence and mutual exclusion information regarding remote accesses. Others have proposed algorithms for analyzing synchronization constructs in the context of framing data flow equations for parallel programs, where strict precedence information is necessary ....

W. W. Carlson and J. M. Draper. Distributed Data Access in AC. In ACM Symposium on Principles and Practice of Parallel Programming, July 1995.

....communication. The current plan is to translate Java AD into standard Java MPI calls, This approach prevents optimizations like those we are implementing in Titanium. Split C. The parallel execution model and global address space support in Titanium are closely related to Split C [5] and AC [3]. Titanium shares a common communication layer with Split C on distributed memory machines, which we have extended as part of the Titanium project to run on shared memory machines. Split C differs from Titanium in that the default pointer type is local rather than global; a local pointer default ....

W.W. Carlson and J.M. Draper. Distributed Data Access in AC. In Proceedings of the 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPOPP), Santa Barbara, CA, July 1995.

....using the primitives that underly the Split C language [7] Split C is an explicitly parallel SPMD language, which provides a global address space abstraction through language features such as global pointers and spread arrays. The global address space features are similar to those in CC [3] AC [2], and other languages. Global pointers, which in Split C can be distinguished from standard pointers through static typing, are wide pointers that represent a processor number and an offset within that processor s address space. Synchronous remote accesses are executed when global pointers are ....

W. Carlson and J. Draper. Distributed Data Access in AC. In Proceedings of the 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPOPP), Santa Barbara, CA, July 1995.

No context found.

W. W. Carlson and J. M. Draper. Distributed data access in AC. In Proc. 5th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP'95, pages 39--47, Santa Barbara, California, July 1995. IDA Supercomputing Research Center.

No context found.

Carlson, William W. and Jesse M. Draper, "Distributed Data Access in AC," Proceedings of the Fifth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPOPP), Santa Barbara, CA, July 1921, 1995, pp. 39-47.

No context found.

Carlson, William W. and Jesse M.Draper, "Distributed Data Access in AC," Proceedings of the Fifth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPOPP), Santa Barbara, CA, July 19-21, 1995, pp.39-47.

No context found.

William W. Carlson and Jesse M. Draper. Distributed data access in AC. In Fifth ACM Sigplan Symposium on Principles and Practices of Parallel Programming, 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