D. K. Chen and P. C. Yew. Statement re-ordering for DOACROSS loops. In International Conference on Parallel Processing, pages 24--28, August 1994. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Compiler Optimization of Scalar Value Communication.. - Zhai, Colohan.. (2002) (5 citations) (Correct)
....forwarding path using intra epoch control speculation and data dependence speculation in Section 4.2. 1. 3 Related Work Parallelization of a loop where the compiler synchronizes a loopcarried data dependence is known as a DOACROSS [8, 26] parallelization and has been exploited in previous work [6, 22, 38]. All schemes for TLS support include some form of DOACROSS synchronization, although few use the compiler to optimize this aspect of speculative execution. The most relevant related work is the Wisconsin Multiscalar [11, 27, 35] compiler, which performs synchronization and scheduling for ....
CHEN, D. K., AND YEW, P. C. Statement re-ordering for DOACROSS loops. In International Conference on Parallel Processing (Aug. 1994), pp. 24--28.
Statement Re-Ordering for DOACROSS Loops - Chen, Yew (1994) (2 citations) Self-citation (Chen Yew) (Correct)
....find SCCs in G Gamma fvg, which always have n Gamma 1 nodes. The largest SCC found in this process is the first PSCC. After the first PSCC is found, it is deleted from G, and Tarjan s algorithm is used to find the remaining PSCCs. In addition, we showed that the condensed graph G will be a ring [5]. PSCC Ordering. The goal of ordering the PSCC is to minimize the distance between the source and the sink PSCCs of any backward dependences under the constraint that no loopindependent dependence (with dependence distance 0) can go backward. There are three cases to consider: Case 1: If there ....
....or (sccb ; scca ; sccc ; sccd) see Figure 4(b) one of the dependences could be covered. 2 On the other hand, if they are ordered as shown in Figure 4(c) no dependence can be covered. Hence, the key here is to avoid the dependenceedge crossings as much as we could. In a corresponding report [5], we prove that the problem 6 b 1 a 2 3 4 d 5 c (c) 6 b 1 a 2 3 4 d 5 c 6 b 1 a 2 3 4 d 5 c 2 3 4 1 5 a c d 6 b (a) b) 6 b 1 a 2 3 4 d 5 c Figure 4: Dependence covering examples. of maximizing the dependence covering is NP complete and a greedy algorithm to ....
D.-K. Chen and P.-C. Yew. Statement re-ordering for DOACROSS loops. CSRD Report No. 1331, Center for Supercomputing Researchand Development, Univ. of Illinois at Urbana-Champaign, January 1994.
Performance Study of a Concurrent Multithreaded Processor - Tsai, Jiang, Ness, Yew (1998) (19 citations) Self-citation (Yew) (Correct)
....overlap of concurrent threads by minimizing the stalls caused by data dependences between threads. To do this, it needs to schedule target stores as early as possible, and schedule load instructions that may be data dependent on the target stores of some predecessor threads as late as possible [1]. The following superthreading specific optimization techniques [14] are also found to be useful: Conversion of Data Speculation to Control Speculation Hardware support for full data speculation can be very expensive, because it needs a buffer (called Address Resolution Buffer in multiscalar [3, ....
D.-K. Chen and P.-C. Yew. Statement reordering for doacross loops. In Proceedings of International Conference on Parallel Processing, volume Vol. II, pages 24--28, August 1994.
Compiler Optimization of Value Communication for Thread-Level.. - Zhai (2005) (Correct)
No context found.
D. K. Chen and P. C. Yew. Statement re-ordering for DOACROSS loops. In International Conference on Parallel Processing, pages 24--28, August 1994.
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