| A. Darte and F. Vivien. Parallelizing Nested Loops with Approximation Distance Vectors: a Survey . Parallel Processing Letters, 7(2):133--144, 1997. |
....size; the change of the problem sizes requires scheduler to re do the entire computation mapping and optimization. To address the aforementioned weakness, this paper proposes a symbolic scheduling algorithm using the parameterized task graph model. The previous work in parallelizing compilers [2, 7, 9] has studied the compact parallelism representation based on finegrain level dependence analysis. The unique aspect of our parameterized task model is that each task node is coarse grained, performing a chunk of computation. The past work has used fine grain dependence information to partition ....
A. Darte and F. Vivien. Parallelizing Nested Loops with Approximation Distance Vectors: a Survey . Parallel Processing Letters, 7(2):133--144, 1997.
....compilers for automatic parallelization (for example SUIF [2] or lastly, task parallelism languages (for example Pyrros [25] Our work is an intermediate approach between the last two ones. In this paper we show how to automatically generate a program from a task graph representation. Compilers [3, 12, 14] are, in general, based on a ne grain representation of the program. Most of the time this implies that communication and computation are considered as unitary. This implies also that all the instructions are modeled, which is very precise but rather complex. Our work is based on a coarse grain ....
Alain Darte and Frederic Vivien. Parallelizing Nested Loops with Approximation Distance Vectors: a Survey . Parallel Processing Letters, 7(2):133-144, 1997.
....use. Two additional surveys have presented the link between loop parallelization algorithms and dependence analysis: in [41] Yang, Ancourt and Irigoin characterize, for each loop transformation used to reveal parallelism, the minimal dependence abstraction needed to check its validity; in [17], a complementary study is proposed that answers the dual question: for a given dependence abstraction, what is the simplest algorithm that detects maximal parallelism Loop parallelization algorithms consist in nding a igoodj loop transformation that reveals parallelism, but we must keep in mind ....
Alain Darte and Fr#d#ric Vivien. Parallelizing nested loops with approximation of distance vectors: a survey. Parallel Processing Letters, 1997.
No context found.
Alain Darte and Frederic Vivien. Parallelizing Nested Loops with Approximation Distance Vectors: a Survey . Parallel Processing Letters, 7(2):133-144, 1997.
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