Schauser, Klaus Erik. Compiling Dataflow into Threads: Efficient Compiler-Controlled Multithreading for Lenient Parallel Languages. Master's Project, Department of Elec- trical Engineering and Computer Sciences, University of California, Berkeley, 1991. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Execution of Dataflow Programs on General-Purpose Hardware - Spertus (1992) (1 citation) (Correct)
....j depends on i, the dependence must not involve an operation of unbounded latency; for example, an instruction that initiates an I structure read must not occur in the same thread as an instruction that uses the returned value. For details on how an Id codeblock is correctly sequentialized, see [19, 26]. When a codeblock is invoked, a frame is allocated for storage of arguments, local variables, entry counts, and a remote continuation vector (RCV) Each entry count indicates the number of times the associated thread must be posted or forked before it may run, representing the number of control ....
....slots into registers where they will remain during the entire schedul ing quantum, at the end of which the exit thread will store them into frame memory. For a fuller description of TAM, see [7] For a partial specification of TL0, see [30] For details on compiling Id programs to run on TAM, see [19]. The reasons for copying the LCV into the RCV are two fold: First, the LCV s maximum size (equal to the number of inlets, each of which may post just one thread) is typically smaller than that of the RCV (equal to the number of threads) so separating the two saves memory. Second, the LCV can be ....
[Article contains additional citation context not shown here]
Schauser, Klaus Erik. Compiling Dataflow into Threads: Efficient Compiler-Controlled Multithreading for Lenient Parallel Languages. Master's Project, Department of Elec- trical Engineering and Computer Sciences, University of California, Berkeley, 1991.
Execution of Dataflow Programs on General-Purpose Hardware - Spertus (1992) (1 citation) (Correct)
....depends on i, the dependence must not involve an operation of unbounded latency; for example, an instruction that initiates an I structure read must not occur in the same 13 thread as an instruction that uses the returned value. For details on how an Id codeblock is correctly sequentialized, see [19, 26]. When a codeblock is invoked, a frame is allocated for storage of arguments, local variables, entry counts, and a remote continuation vector (RCV) Each entry count indicates the number of times the associated thread must be posted or forked before it may run, representing the number of control ....
....slots into registers where they will remain during the entire scheduling quantum, at the end of which the exit thread will store them into frame memory. For a fuller description of TAM, see [7] For a partial specification of TL0, see [30] For details on compiling Id programs to run on TAM, see [19]. 1 The reasons for copying the LCV into the RCV are two fold: First, the LCV s maximum size (equal to the number of inlets, each of which may post just one thread) is typically smaller than that of the RCV (equal to the number of threads) so separating the two saves memory. Second, the LCV can ....
[Article contains additional citation context not shown here]
Schauser, Klaus Erik. Compiling Dataflow into Threads: Efficient Compiler-Controlled Multithreading for Lenient Parallel Languages. Master's Project, Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, 1991.
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