Browsing by Subject "Sparse matrix-vector multiplication"
Now showing 1 - 8 of 8
Results Per Page
Sort Options
Item Open Access Addressing volume and latency overheads in 1d-parallel sparse matrix-vector multiplication(Springer, 2017-08-09) Acer, Seher; Selvitopi, Oğuz; Aykanat, CevdetThe scalability of sparse matrix-vector multiplication (SpMV) on distributed memory systems depends on multiple factors that involve different communication cost metrics. The irregular sparsity pattern of the coefficient matrix manifests itself as high bandwidth (total and/or maximum volume) and/or high latency (total and/or maximum message count) overhead. In this work, we propose a hypergraph partitioning model which combines two earlier models for one-dimensional partitioning, one addressing total and maximum volume, and the other one addressing total volume and total message count. Our model relies on the recursive bipartitioning paradigm and simultaneously addresses three cost metrics in a single partitioning phase in order to reduce volume and latency overheads. We demonstrate the validity of our model on a large dataset that contains more than 300 matrices. The results indicate that compared to the earlier models, our model significantly improves the scalability of SpMV. © 2017, Springer International Publishing AG.Item Open Access Locality-aware parallel sparse matrix-vector and matrix-transpose-vector multiplication on many-core processors(Institute of Electrical and Electronics Engineers, 2016) Karsavuran, M. O.; Akbudak K.; Aykanat, CevdetSparse matrix-vector and matrix-transpose-vector multiplication (SpMMTV) repeatedly performed as z ← ATx and y ← A z (or y ← A w) for the same sparse matrix A is a kernel operation widely used in various iterative solvers. One important optimization for serial SpMMTV is reusing A-matrix nonzeros, which halves the memory bandwidth requirement. However, thread-level parallelization of SpMMTV that reuses A-matrix nonzeros necessitates concurrent writes to the same output-vector entries. These concurrent writes can be handled in two ways: via atomic updates or thread-local temporary output vectors that will undergo a reduction operation, both of which are not efficient or scalable on processors with many cores and complicated cache-coherency protocols. In this work, we identify five quality criteria for efficient and scalable thread-level parallelization of SpMMTV that utilizes one-dimensional (1D) matrix partitioning. We also propose two locality-aware 1D partitioning methods, which achieve reusing A-matrix nonzeros and intermediate z-vector entries; exploiting locality in accessing x -, y -, and -vector entries; and reducing the number of concurrent writes to the same output-vector entries. These two methods utilize rowwise and columnwise singly bordered block-diagonal (SB) forms of A. We evaluate the validity of our methods on a wide range of sparse matrices. Experiments on the 60-core cache-coherent Intel Xeon Phi processor show the validity of the identified quality criteria and the validity of the proposed methods in practice. The results also show that the performance improvement from reusing A-matrix nonzeros compensates for the overhead of concurrent writes through the proposed SB-based methods.Item Open Access A novel method for scaling iterative solvers: avoiding latency overhead of parallel sparse-matrix vector multiplies(Institute of Electrical and Electronics Engineers, 2015) Selvitopi, R. O.; Ozdal, M. M.; Aykanat, CevdetIn parallel linear iterative solvers, sparse matrix vector multiplication (SpMxV) incurs irregular point-to-point (P2P) communications, whereas inner product computations incur regular collective communications. These P2P communications cause an additional synchronization point with relatively high message latency costs due to small message sizes. In these solvers, each SpMxV is usually followed by an inner product computation that involves the output vector of SpMxV. Here, we exploit this property to propose a novel parallelization method that avoids the latency costs and synchronization overhead of P2P communications. Our method involves a computational and a communication rearrangement scheme. The computational rearrangement provides an alternative method for forming input vector of SpMxV and allows P2P and collective communications to be performed in a single phase. The communication rearrangement realizes this opportunity by embedding P2P communications into global collective communication operations. The proposed method grants a certain value on the maximum number of messages communicated regardless of the sparsity pattern of the matrix. The downside, however, is the increased message volume and the negligible redundant computation. We favor reducing the message latency costs at the expense of increasing message volume. Yet, we propose two iterative-improvement-based heuristics to alleviate the increase in the volume through one-to-one task-to-processor mapping. Our experiments on two supercomputers, Cray XE6 and IBM BlueGene/Q, up to 2,048 processors show that the proposed parallelization method exhibits superior scalable performance compared to the conventional parallelization method.Item Open Access Optimizing nonzero-based sparse matrix partitioning models via reducing latency(Academic Press, 2018) Acer, S.; Selvitopi, O.; Aykanat, CevdetFor the parallelization of sparse matrix-vector multiplication (SpMV) on distributed memory systems, nonzero-based fine-grain and medium-grain partitioning models attain the lowest communication volume and computational imbalance among all partitioning models. This usually comes, however, at the expense of high message count, i.e., high latency overhead. This work addresses this shortcoming by proposing new fine-grain and medium-grain models that are able to minimize communication volume and message count in a single partitioning phase. The new models utilize message nets in order to encapsulate the minimization of total message count. We further fine-tune these models by proposing delayed addition and thresholding for message nets in order to establish a trade-off between the conflicting objectives of minimizing communication volume and message count. The experiments on an extensive dataset of nearly one thousand matrices show that the proposed models improve the total message count of the original nonzero-based models by up to 27% on the average, which is reflected on the parallel runtime of SpMV as an average reduction of 15% on 512 processors.Item Open Access Reduce operations: send volume balancing while minimizing latency(IEEE, 2020) Karsavuran, M. Ozan; Acer, S.; Aykanat, CevdetCommunication hypergraph model was proposed in a two-phase setting for encapsulating multiple communication cost metrics (bandwidth and latency), which are proven to be important in parallelizing irregular applications. In the first phase, computational-task-to-processor assignment is performed with the objective of minimizing total volume while maintaining computational load balance. In the second phase, communication-task-to-processor assignment is performed with the objective of minimizing total number of messages while maintaining communication-volume balance. The reduce-communication hypergraph model suffers from failing to correctly encapsulate send-volume balancing. We propose a novel vertex weighting scheme that enables part weights to correctly encode send-volume loads of processors for send-volume balancing. The model also suffers from increasing the total communication volume during partitioning. To decrease this increase, we propose a method that utilizes the recursive bipartitioning framework and refines each bipartition by vertex swaps. For performance evaluation, we consider column-parallel SpMV, which is one of the most widely known applications in which the reduce-task assignment problem arises. Extensive experiments on 313 matrices show that, compared to the existing model, the proposed models achieve considerable improvements in all communication cost metrics. These improvements lead to an average decrease of 30 percent in parallel SpMV time on 512 processors for 70 matrices with high irregularity.Item Open Access Reducing latency cost in 2D sparse matrix partitioning models(Elsevier BV, 2016) Selvitopi, O.; Aykanat, CevdetSparse matrix partitioning is a common technique used for improving performance of parallel linear iterative solvers. Compared to solvers used for symmetric linear systems, solvers for nonsymmetric systems offer more potential for addressing different multiple communication metrics due to the flexibility of adopting different partitions on the input and output vectors of sparse matrix-vector multiplication operations. In this regard, there exist works based on one-dimensional (1D) and two-dimensional (2D) fine-grain partitioning models that effectively address both bandwidth and latency costs in nonsymmetric solvers. In this work, we propose two new models based on 2D checkerboard and jagged partitioning. These models aim at minimizing total message count while maintaining a balance on communication volume loads of processors; hence, they address both bandwidth and latency costs. We evaluate all partitioning models on two nonsymmetric system solvers implemented using the widely adopted PETSc toolkit and conduct extensive experiments using these solvers on a modern system (a BlueGene/Q machine) successfully scaling them up to 8K processors. Along with the proposed models, we put practical aspects of eight evaluated models (two 1D- and six 2D-based) under thorough analysis. To the best of our knowledge, this is the first work that analyzes practical performance of 2D models on this scale. Among evaluated models, the models that rely on 2D jagged partitioning obtain the most promising results by striking a balance between minimizing bandwidth and latency costs.Item Open Access Reordering methods for exploiting spatial and temporal localities in parallel sparse matrix-vector multiplication(2016-08) AbuBaker, NabilSparse Matrix-Vector multiplication (SpMV) is a very important kernel operation for many scientific applications. For irregular sparse matrices, the SpMV operation suffers from poor cache performance due to the irregular accesses of the input vector entries. In this work, we propose row and column reordering methods based on Graph partitioning (GP) and Hypergraph partitioning (HP) in order to exploit spatial and temporal localities in accessing input vector entries by clustering rows/columns with a similar sparsity pattern close to each other. The proposed methods exploit spatial and temporal localities separately (using either rows or columns of the matrix in a GP or HP method), simultaneously (using both rows and column) and in a two-phased manner(using either rows or columns in each phase). We evaluate the validity of the proposed models on a 60- core Xeon Phi co-processor for a large set of sparse matrices arising from different applications. The performance results confirm the validity and the effectiveness of the proposed methods and models.Item Open Access Spatiotemporal graph and hypergraph partitioning models for sparse matrix-vector multiplication on many-core architectures(IEEE Computer Society, 2019) Abubaker, Nabil; Akbudak, K.; Aykanat, CevdetThere exist graph/hypergraph partitioning-based row/column reordering methods for encoding either spatial or temporal locality for sparse matrix-vector multiplication (SpMV) operations. Spatial and temporal hypergraph models in these methods are extended to encapsulate both spatial and temporal localities based on cut/uncut net categorization obtained from vertex partitioning. These extensions of spatial and temporal hypergraph models encode the spatial locality primarily and the temporal locality secondarily, and vice-versa, respectively. However, the literature lacks models that simultaneously encode both spatial and temporal localities utilizing only vertex partitioning for further improving the performance of SpMV on shared-memory architectures. In order to fill this gap, we propose a novel spatiotemporal hypergraph model that leads to a one-phase spatiotemporal reordering method which encodes both types of locality simultaneously. We also propose a framework for spatiotemporal methods which encodes both types of locality in two dependent phases and two separate phases. The validity of the proposed spatiotemporal models and methods are tested on a wide range of sparse matrices and the experiments are performed on both a 60-core Intel Xeon Phi processor and a Xeon processor. Results show the validity of the methods via almost doubling the Gflop/s performance through enhancing data locality in parallel SpMV operations.