Buhr, Peter A. and Stroobosscher, R. A. The uSystem: providing light-weight concurrency on shared memory multiprocessor computers running UNIX. Software---Practice and Experience 20, 9 (Sept. 1990), 929-964. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Filaments: Efficient Support for Fine-Grain Parallelism - Engler, Andrews, Lowenthal (1994) (21 citations) (Correct)
....Third, the Filaments package cannot support any application that requires that a thread be able to block, e.g. a multithreaded event handler or a file server. An easy way to get around the last two problems is simply to install Filaments on top of a widely supported thread package (e.g. uSystem [Buhr88]) and to use its threads as the Filaments servers. Unfortunately, the only ways we see to solve the code size problem are to revert back to contexts or to resort to runtime code generation. 5. Related Work There is a wealth of related research on threads packages, some of which support ....
....or to resort to runtime code generation. 5. Related Work There is a wealth of related research on threads packages, some of which support fine grain parallelism. The first step towards efficient parallelism was lightweight thread packages such as Threads [Doep87] Presto [Bers88] System [Buhr88], C Threads [Coop90] C [Buhr92] and Sun Lightweight Processes [Stei92] We will call these standard packages to distinguish them from threads packages that do not have a stack for each thread. The goal of standard packages is to provide the user with a natural thread abstraction, and many of ....
Buhr, Peter A. and Stroobosscher, R. A. The uSystem: providing light-weight concurrency on shared memory multiprocessor computers running UNIX. Software---Practice and Experience 20, 9 (Sept. 1990), 929-964.
Architecture-Independent Parallelism for Both Sharedand.. - Lowenthal, Freeh (Correct)
....of efficient threads packages, distributed shared memory, and overlapping communication and computation. Furthermore, it is one of many approaches to architecture independence. 5. 1 Threads There are many general purpose thread packages, including Threads [Doe87] Presto [BLL88] System [BS90] C [BDSY92] and Sun Lightweight Processes [SS92] All of the above packages support pre emption to provide fairness, which requires each thread to maintain a private stack. Consequently, the package has to perform a full context switch when it switches between threads, which makes efficient ....
Peter A. Buhr and R.A. Stroobosscher. The uSystem: providing light-weight concurrency on shared memory multiprocessor computers running UNIX. Software Practice and Experience, pages 929--964, September 1990.
Efficient Support for Fine-Grain Parallelism on.. - Lowenthal, Freeh.. (1999) (2 citations) (Correct)
....because the Symmetry is now an obsolete machine, we cannot repeat the experiments on the new package. 5 Related Work There are many existing threads packages. The initial attempts to support efficient parallelism were lightweight thread packages such as Threads [Doe87] Presto [BLL88] System [BS90], C [BDSY92] and Sun Lightweight Processes [SS92] We will call these standard packages because they have a stack for each thread. The goal of standard packages is to provide the user with a natural thread abstraction and many of the usual concurrent programming primitives; different packages ....
Peter A. Buhr and R.A. Stroobosscher. The uSystem: providing light-weight concurrency on shared memory multiprocessor computers running UNIX. Software Practice and Experience, pages 929--964, September 1990.
Distributed Filaments: Efficient Fine-Grain Parallelism.. - Freeh, Lowenthal.. (1994) (46 citations) (Correct)
....shown in Figure 12. Overlapping communication leads to a 9 and 21 improvement on 4 and 8 nodes, as can be seen by comparing the times in Figure 12 with those in Figure 5. 5 Related Work There is a wealth of related work on threads packages. In general purpose threads packages, such as uSystem [BS90] each thread has its own context. Consequently, the package has to perform a full context switch when it switches between threads. This is true even in threads packages with Nodes 1 2 4 8 DF Time (sec) 212 103 61.4 40.9 DF Speedup 1.01 2.09 3.50 5.26 Figure 11: Jacobi iteration, ....
Peter A. Buhr and R.A. Stroobosscher. The uSystem: providing light-weight concurrency on shared memory multiprocessor computers running UNIX. Software Practice and Experience, pages 929--964, September 1990.
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