Parallel Alternating-Direction Access Machine

Parallel Alternating - Direction Access Machine ? Bogdan

This paper presents a theoretical study of a model of parallel computations called Parallel Alternating-Direction Access Machine (padam). padam is an abstraction of the multiprocessor computers adena /adenart and a prototype architecture usc/omp. The main feature of padam is the organization of access to the global memory: (1) the memory modules are arranged as a 2-dimensional array, (2) each p...

Algorithms for the parallel alternating direction access machine

We describe a number of algorithms for the model for parallel computation called Parallel Alternating-Direction Access Machine (padam). This model has the memory modules of the global memory arranged as a 2-dimensional array, with each processor assigned to a row and a column, the processors can switch synchronously between row and column access modes. We study the issues of inter-processor com...

Linearized Alternating Direction Method with Parallel Splitting and Adaptive Penalty for Separable Convex Programs in Machine Learning – Supplementary Materials

A Performance Estimator for Parallel Programs

In this paper we describe a Parallel Performance Estimator suitable for the comparative evaluation of parallel algorithms. The Esti-mator is designed for SPMD programs written in either C or FORTRAN. Simulation is used to produce estimates of execution times for varying numbers of processors and to analyse the communication overheads. Results from the estimator are compared with actual results ...

Efficient Distributed Linear Classification Algorithms via the Alternating Direction Method of Multipliers

Linear classification has demonstrated success in many areas of applications. Modern algorithms for linear classification can train reasonably good models while going through the data in only tens of rounds. However, large data often does not fit in the memory of a single machine, which makes the bottleneck in large-scale learning the disk I/O, not the CPU. Following this observation, Yu et al....