Titlebook: Languages and Compilers for Parallel Computing; 21th International W José Nelson Amaral Conference proceedings 2008 Springer-Verlag Berlin

Pandemic

Minimum Lock Assignment: A Method for Exploiting Concurrency among Critical Sections, the optimality of the heuristic with the Integer Linear Programming (ILP) solver. We have also tested the efficiency of the heuristic using scientific applications, from which we obtain up to 30% performance gain with respect to the programs in which all critical sections are controlled by a single lock.

反复拉紧

Set-Congruence Dynamic Analysis for Thread-Level Speculation (TLS),r congruences, which can be employed on-line or off-line. One, ., efficiently enables loop iterations to be allocated to threads (and calculates the maximum effective concurrency); the other, ., calculates a hash function that reduces space overhead and performance loss due to TLS-metadata-based and cache-based task interference.

famine

Thread Safety through Partitions and Effect Agreements, flow-sensitive effect system requires methods to disclose which partitions of the heap they will read or write, and also allows them to specify an . which can be used to limit the conditions in which a method can be called.

压倒

Conference proceedings 2008ollowing its long-established tradition, the workshop focused on topics at the frontierofresearchanddevelopmentinlanguages,optimizingcompilers,appli- tions, and programming models for high-performance computing. While LCPC continues to focus on parallel computing, the 2008 edition included the pres-

一条卷发

CESS

978-3-540-89739-2Springer-Verlag Berlin Heidelberg 2008

CREEK

Languages and Compilers for Parallel Computing978-3-540-89740-8Series ISSN 0302-9743 Series E-ISSN 1611-3349

Baffle

分贝

avenge

Automatic Pre-Fetch and Modulo Scheduling Transformations for the Cell BE Architecture,tically eliminate the software cache overhead in the innermost loop. The cache design enables automatic pre-fetch and modulo scheduling transforma-tions. Performance evaluation indicates that the optimized software-cache structures combined with the proposed pre-fetch techniques translate into speed