| 1. |
- Grahn, Håkan, et al.
(författare)
-
Evaluation of a Competitive-Update Cache Coherence Protocol with Migratory Data Detection
- 1996
-
Ingår i: Journal of Parallel and Distributed Computing. - San Diego : Academic. - 0743-7315 .- 1096-0848. ; 39:2, s. 168-180
-
Tidskriftsartikel (refereegranskat)abstract
- Although directory-based write-invalidate cache coherence protocols have a potential to improve the performance of large-scale multiprocessors, coherence misses limit the processor utilization. Therefore, so-called competitive-update protocols-hybrid protocols that on a per-block basis dynamically switch between write-invalidate and write-update-have been considered as a means to reduce the coherence miss rate and have been shown to be a better coherence policy for a wide range of applications. Unfortunately, such protocols may cause high traffic peaks for applications with extensive use of migratory objects. These traffic peaks can offset the performance gain of a reduced miss rate if the network bandwidth is not sufficient. We propose in this study to extend a competitive-update protocol with a previously published adaptive mechanism that can dynamically detect migratory objects and reduce the coherence traffic they cause. Detailed architectural simulations based on five scientific and engineering applications show that this adaptive protocol outperforms a write-invalidate protocol by reducing the miss rate and bandwidth needed by up to 71 and 26%, respectively.
|
|
| 2. |
- Lundberg, Lars
(författare)
-
Predicting and bounding the speedup of multithreaded Solaris programs
- 1999
-
Ingår i: Journal of Parallel and Distributed Computing. - SAN DIEGO : ACADEMIC PRESS INC. - 0743-7315 .- 1096-0848. ; , s. 322-333
-
Tidskriftsartikel (refereegranskat)abstract
- In Solaris, threads are frequently relocated. The data associated with a relocated thread have to be moved from the cache of the old processor to the new processor. In order to avoid poor memory performance due to thread relocation, threads can be bound to processors-static scheduling. Finding a static schedule which results in maximum speedup is NP-hard. It is even difficult to determine if a static schedule is close to the optimal case or not. Here, a technique for predicting the speedup of multithreaded Solaris programs is presented. Based on an existing theoretical result, a lower bound on the maximal speedup is also obtained. The predicted speedup and the bound are based on recordings from a single-processor execution. When comparing the predictions with the real speedup using a multiprocessor with eight processors, we see that the predictions are very good. By comparing the speedup of a static schedule with the bound, we see that it is worthwhile to look for other schedules. (C) 1999 Academic Press.
|
|
| 3. |
- Stoyenko, AD, et al.
(författare)
-
Load balanced mapping of distributed objects to minimize network communication
- 1996
-
Ingår i: Journal of Parallel and Distributed Computing. - SAN DIEGO : ACADEMIC PRESS INC JNL-COMP SUBSCRIPTIONS. - 0743-7315 .- 1096-0848. ; , s. 117-136
-
Tidskriftsartikel (refereegranskat)abstract
- This paper introduces a new load balancing and communication minimizing heuristic used in the In verse Remote Procedure Call (IRPC) system. While the paper briefly describes the IRPC system, the focus is on the new IRPC assignment heuristic. The IRPC compiler maps a distributed program to a graph that represents program objects and their dependencies (due to invocations and parameter passing) as nodes and edges, respectively. In the graph, the system preserves conditional and iterative flows, records network transmission and execution costs, and marks nodes that have to reside at specific network sites. The graph is then partitioned by the heuristic to derive a (sub)optimal node assignment to network sites minimizing load balancing and network data transport. The resulting program partition is then reflected in the physical object distribution, and remote and local object communication is transparently implemented. The compiler and run-time system use efficient implementation techniques such as type prediction, inlining, splitting and subprogram passing. The last of these allows remote code to be copied to local data, as an alternative to copying data to the remote site, whenever this will reduce network data transport. The IRPC graph partitioning heuristic operates in time O(E(log d + l + log M)), where M is the number of network sites, E is the number of communication edges, and d is the maximum degree of a node; l is a parameter of the algorithm, and can vary between 1 and N, where N is the number of communicating objects. This complexity is more nearly independent of M, and considerably better in terms of E and N, than that of previously known related algorithms, such as A*, which employs backtracking and is potentially exponential, or the max-flow/min-cut class of network flow algorithms or heuristics which tend to be at least of Omega(MN(2)E), and it can be made (by choosing l appropriately) as efficient as even such fast heuristics as heaviest-edge-first, minimal communication, and Kernighan-Lin. In an extensive quantitative evaluation, the heuristic has been demonstrated to perform very well, giving on the average 75% traffic cost reductions for over 95% of the programs when compared to random partitioning, and outperforming in cost reduction and actual execution time the three aforementioned fast heuristics, even with a large l. Thus, to the best of our knowledge, this is the first report of a well-performing assignment heuristic that is both essentially linear in the number of communication edges, and better than existing, established heuristics of no better complexity. (C) 1996 Academic Press, Inc.
|
|
| 4. |
- Veanes, Margus, et al.
(författare)
-
Natural cycletrees : Flexible interconnection graphs
- 1996
-
Ingår i: Journal of Parallel and Distributed Computing. - : Elsevier BV. - 0743-7315 .- 1096-0848. ; 33:1, s. 44-54
-
Tidskriftsartikel (refereegranskat)abstract
- Natural cycletrees, formally defined in this paper, is a subclass of Hamiltonian graphs with maximum degree 3 that contain a binary spanning tree. A natural cycletree used as an interconnection network thus supports directly broadcasting through the binary tree as well as nearest-neighbor communication through the cycle. Natural cycletrees have several other interesting properties; e.g., they are planar, easily extensible, and can be contracted using the same methods as for binary trees. The main results of the paper are: (i) Given an arbitrary basic binary spanning treeT, there exists a natural cycletree with a minimal number of edges forT. (ii) A natural cycletree has a very simple router. We give a superfast parallel algorithm that can establish near optimal router data for that router.
|
|
