The system of linear equations obtained by finite difference discretization of the three-dimensional Poisson problem is solved by a matrix-free parallel implementation of the Jacobi method. After checking the convergence of our implementation, we compare the speed of using blocking MPI_Ssend/MPI_Recv and non-blocking MPI_Isend/MPI_Irecv communications. The comparisons indicate that the non-bloc...

We present an efficient and practical lock-free implementation of a concurrent priority queue that is suitable for both fully concurrent (large multi-processor) systems as well as pre-emptive (multi-process) systems. Many algorithms for concurrent priority queues are based on mutual exclusion. However, mutual exclusion causes blocking which has several drawbacks and degrades the system’s overal...

The design of a FIFO buuer with fully asynchronous, non-blocking write and read operation is presented. The buuer's logic consists of a standard self-timed FIFO (e.g. pipeline) surrounded by two special interface blocks, In and Out, connected to Writer and Reader, respectively. To avoid temporal blocking of Writer requesting for a write when the pipeline is full the In block implements skipping...

Mobile computing raises many new issues, such as lack of stable storage, low bandwidth of wireless channel, high mobility, and limited battery life. These new issues make traditional checkpointing algorithms unsuitable. Prakash and Singhal [14] proposed the first coordinated checkpointing algorithm for mobile computing systems. However, we showed that their algorithm may result in an inconsiste...

Nowadays, clusters and grids are made of more and more computing nodes. The programming of multi-processes applications is the most often achieved through message passing. The increase of the number of processes implies that theses applications need to use a fault tolerant message passing library. In this paper, we present two implementations of fault tolerant protocols based on MPICH, a blocki...

Shared objects are used in the distributed environment for synchronization, date sharing and service provision. They are usually replicated in a distributed system for higher availability and fault tolerance. Coherent or strongly consistent implementation of replicated objects has been studied extensively in the literature. This paper concentrates on a general method for specifying and implemen...

We study the π-calculus, enriched with pairing and non-blocking input, and define a notion of type assignment that uses the type constructor →. We encode the circuits of the calculus X into this variant of π, and show that all reduction (cut-elimination) and assignable types are preserved. Since X enjoys the Curry-Howard isomorphism for Gentzen’s calculus LK, this implies that all proofs in LK ...

Path allocation in a three-stage ATM switch is the process whereby bandwidth is reserved through the second stage of the switch for each cell. Cell-level path allocation, performed once in every time slot, ensures that cells are routed through the second stage of the switch without delay or contention. A new algorithm for cell-level path allocation was recently proposed by the author. The motiv...

Most published concurrent data structures which avoid locking do not provide any fairness guarantees. That is, they allow processes to access a data structure and complete their operations arbitrarily many times before some other trying process can complete a single operation. Such a behavior can be prevented by enforcing fairness. However, fairness requires waiting or helping. Helping techniqu...

This paper presents a fault-tolerant lazy replication protocol that ensures 1-copy serializability at a relatively low cost. Unlike eager replication approaches, our protocol enables local transaction execution and does not lead to any deadlock situation. Compared to previous lazy replication approaches, we significantly reduce the abort rate of transactions and we do not require any reconcilia...