perm filename ASD[VLI,LSP] blob
sn#381934 filedate 1978-09-08 generic text, type T, neo UTF8
o the increased capacity for parallelism.
11590 However, the performance improvement is mitigated by the additional
11595 overheads involved in manipulating the new work units.
11600
11605 Beacuse the environment is dynamically changing, the allocation of
11610 work unitsin a task force of cooperating processes must also be
11615 performed dynamically. Failure to do so prevents the processes from
11620 cooperating to compensate for the computational changes that result
11625 during the execution of the program.
11630
11635 Finally, task forces arranged asynchronously permit greater utilization
11640 of the available processing power, hence better performance results.
@device<xgp>
@make<report>
@blankspace<18 lines>
@Center<THE IMPLEMENTATION AND EVALUATION OF A PARALLEL ALGORITHM>
@center<ON C.MMP>
@blankspace(4 lines)
@Center<P.N. Oleinick and S.H. Fuller>
@blankspace(4 lines)
@center(Computer Science Department)
@center(Carnegie-Mellon University)
@center(June 6, 1978)
@blankspace(4 lines)
@i<Keywords>: performance evaluation, multiprocessors, synchronization,
parallel algorithms, cooperating processes.
@blankspace(2 lines)
The research described here was supported by the Defense Advanced
Research Projects Agency (Contract: F44620-73-C-0074, monitored by
the Air Force Office of Scientific Research), and in part by the
Office of Naval Research (Contract: N00014-77-C-0500).