Browsing by Subject "Linear algebra"

Now showing 1 - 3 of 3

Open Access
Improving performance of sparse matrix dense matrix multiplication on large-scale parallel systems
(Elsevier BV, 2016) Acer, S.; Selvitopi, O.; Aykanat, Cevdet
We propose a comprehensive and generic framework to minimize multiple and different volume-based communication cost metrics for sparse matrix dense matrix multiplication (SpMM). SpMM is an important kernel that finds application in computational linear algebra and big data analytics. On distributed memory systems, this kernel is usually characterized with its high communication volume requirements. Our approach targets irregularly sparse matrices and is based on both graph and hypergraph partitioning models that rely on the widely adopted recursive bipartitioning paradigm. The proposed models are lightweight, portable (can be realized using any graph and hypergraph partitioning tool) and can simultaneously optimize different cost metrics besides total volume, such as maximum send/receive volume, maximum sum of send and receive volumes, etc., in a single partitioning phase. They allow one to define and optimize as many custom volume-based metrics as desired through a flexible formulation. The experiments on a wide range of about thousand matrices show that the proposed models drastically reduce the maximum communication volume compared to the standard partitioning models that only address the minimization of total volume. The improvements obtained on volume-based partition quality metrics using our models are validated with parallel SpMM as well as parallel multi-source BFS experiments on two large-scale systems. For parallel SpMM, compared to the standard partitioning models, our graph and hypergraph partitioning models respectively achieve reductions of 14% and 22% in runtime, on average. Compared to the state-of-the-art partitioner UMPa, our graph model is overall 14.5 ï¿½ faster and achieves an average improvement of 19% in the partition quality on instances that are bounded by maximum volume. For parallel BFS, we show on graphs with more than a billion edges that the scalability can significantly be improved with our models compared to a recently proposed two-dimensional partitioning model.
Open Access
LinGraph: a graph-based automated planner for concurrent task planning based on linear logic
(Springer New York LLC, 2017) Kortik S.; Saranli, U.
In this paper, we introduce an automated planner for deterministic, concurrent domains, formulated as a graph-based theorem prover for a propositional fragment of intuitionistic linear logic, relying on the previously established connection between intuitionistic linear logic and planning problems. The new graph-based theorem prover we introduce improves planning performance by reducing proof permutations that are irrelevant to planning problems particularly in the presence of large numbers of objects and agents with identical properties (e.g. robots within a swarm, or parts in a large factory). We first present our graph-based automated planner, the Linear Logic Graph Planner (LinGraph). Subsequently we illustrate its application for planning within a concurrent manufacturing domain and provide comparisons with four existing automated planners, BlackBox, Symba-2, Metis and the Temporal Fast Downward (TFD), covering a wide range of state-of-the-art automated planning techniques and implementations. We show that even though LinGraph does not rely on any heuristics, it still outperforms these systems for concurrent domains with large numbers of identical objects and agents. These gains persist even when existing methods on symmetry reduction and numerical fluents are used, with LinGraph capable of handling problems with thousands of objects. Following these results, we also show that plan construction with LinGraph is equivalent to multiset rewriting systems, formally relating LinGraph to intuitionistic linear logic. © 2017, Springer Science+Business Media New York.
Open Access
New characterizations of ℓ1 solutions to overdetermined systems of linear equations
(Elsevier, 1994) Madsen, K.; Nielsen H. B.; Pınar, M. Ç.
New characterizations of the ℓ1 solutions to overdetermined system of linear equations are given. The first is a polyhedral characterization of the solution set in terms of a special sign vector using a simple property of the ℓ1 solutions. The second characterization is based on a smooth approximation of the ℓ1 function using a "Huber" function. This allows a description of the solution set of the ℓ1 problem from any solution to the approximating problem for sufficiently small positive values of an approximation parameter. A sign approximation property of the Huber problem is also considered and a characterization of this property is given. © 1994.