Graph partitioning has emerged as an area of interest due to its use in various applications in computational research. One way to partition a graph is to solve for the eigenvectors of the corresponding graph Laplacian matrix. This project focuses on the eigensolver LOBPCG and the evaluation of a new preconditioner: Randomized Cholesky Factorization (rchol). This proconditioner was tested for its speed and accuracy against other well-known preconditioners for the method. After experiments were run on several known test matrices, rchol appears to be a better preconditioner for structured matrices. This research was sponsored by National Nuclear Security Administration Minority Serving Institutions Internship Program (NNSA-MSIIP) and completed at host facility Sandia National Laboratories. As such, after discussion of the research project itself, this report contains a brief reflection on experience gained as a result of participating in the NNSA-MSIIP.
A graph is a mathematical representation of a network; we say it consists of a set of vertices, which are connected by edges. Graphs have numerous applications in various fields, as they can model all sorts of connections, processes, or relations. For example, graphs can model intricate transit systems or the human nervous system. However, graphs that are large or complicated become difficult to analyze. This is why there is an increased interest in the area of graph partitioning, reducing the size of the graph into multiple partitions. For example, partitions of a graph representing a social network might help identify clusters of friends or colleagues. Graph partitioning is also a widely used approach to load balancing in parallel computing. The partitioning of a graph is extremely useful to decompose the graph into smaller parts and allow for easier analysis. There are different ways to solve graph partitioning problems. For this work, we focus on a spectral partitioning method which forms a partition based upon the eigenvectors of the graph Laplacian (details presented in Acer, et. al.). This method uses the LOBPCG algorithm to compute these eigenvectors. LOBPCG can be accelerated by an operator called a preconditioner. For this internship, we evaluate a randomized Cholesky (rchol) preconditioner for its effectiveness on graph partitioning problems with LOBPCG. We compare it with two standard preconditioners: Jacobi and Incomplete Cholesky (ichol). This research was conducted from August to December 2021 in conjunction with Sandia National Laboratories.
Graph partitioning has been an important tool to partition the work among several processors to minimize the communication cost and balance the workload. While accelerator-based supercomputers are emerging to be the standard, the use of graph partitioning becomes even more important as applications are rapidly moving to these architectures. However, there is no distributed-memory-parallel, multi-GPU graph partitioner available for applications. We developed a spectral graph partitioner, Sphynx, using the portable, accelerator-friendly stack of the Trilinos framework. In Sphynx, we allow using different preconditioners and exploit their unique advantages. We use Sphynx to systematically evaluate the various algorithmic choices in spectral partitioning with a focus on the GPU performance. We perform those evaluations on two distinct classes of graphs: regular (such as meshes, matrices from finite element methods) and irregular (such as social networks and web graphs), and show that different settings and preconditioners are needed for these graph classes. The experimental results on the Summit supercomputer show that Sphynx is the fastest alternative on irregular graphs in an application-friendly setting and obtains a partitioning quality close to ParMETIS on regular graphs. When compared to nvGRAPH on a single GPU, Sphynx is faster and obtains better balance and better quality partitions. Sphynx provides a good and robust partitioning method across a wide range of graphs for applications looking for a GPU-based partitioner.
Abdelfattah, Ahmad A.; Anzt, Hartwig A.; Ayala, Alan A.; Boman, Erik G.; Carson, Erin C.; Cayrols, Sebastien C.; Cojean, Terry C.; Dongarra, Jack D.; Falgout, Rob F.; Gates, Mark G.; Gr\"{u}tzmacher, Thomas G.; Higham, Nicholas J.; Kruger, Scott E.; Li, Sherry L.; Lindquist, Neil L.; Liu, Yang L.; Loe, Jennifer A.; Nayak, Pratik N.; Osei-Kuffuor, Daniel O.; Pranesh, Sri P.; Rajamanickam, Sivasankaran R.; Ribizel, Tobias R.; Smith, Bryce B.; Swirydowicz, Kasia S.; Thomas, Stephen T.; Tomov, Stanimire T.; M. Tsai, Yaohung M.; Yamazaki, Ichitaro Y.;
Yang, Urike M.
Over the last year, the ECP xSDK-multiprecision effort has made tremendous progress in developing and deploying new mixed precision technology and customizing the algorithms for the hardware deployed in the ECP flagship supercomputers. The effort also has succeeded in creating a cross-laboratory community of scientists interested in mixed precision technology and now working together in deploying this technology for ECP applications. In this report, we highlight some of the most promising and impactful achievements of the last year. Among the highlights we present are: Mixed precision IR using a dense LU factorization and achieving a 1.8× speedup on Spock; results and strategies for mixed precision IR using a sparse LU factorization; a mixed precision eigenvalue solver; Mixed Precision GMRES-IR being deployed in Trilinos, and achieving a speedup of 1.4× over standard GMRES; compressed Basis (CB) GMRES being deployed in Ginkgo and achieving an average 1.4× speedup over standard GMRES; preparing hypre for mixed precision execution; mixed precision sparse approximate inverse preconditioners achieving an average speedup of 1.2×; and detailed description of the memory accessor separating the arithmetic precision from the memory precision, and enabling memory-bound low precision BLAS 1/2 operations to increase the accuracy by using high precision in the computations without degrading the performance. We emphasize that many of the highlights presented here have also been submitted to peer-reviewed journals or established conferences, and are under peer-review or have already been published.
Support for lower precision computation is becoming more common in accelerator hardware due to lower power usage, reduced data movement and increased computational performance. However, computational science and engineering (CSE) problems require double precision accuracy in several domains. This conflict between hardware trends and application needs has resulted in a need for multiprecision strategies at the linear algebra algorithms level if we want to exploit the hardware to its full potential while meeting the accuracy requirements. In this paper, we focus on preconditioned sparse iterative linear solvers, a key kernel in several CSE applications. We present a study of multiprecision strategies for accelerating this kernel on GPUs. We seek the best methods for incorporating multiple precisions into the GMRES linear solver; these include iterative refinement and parallelizable preconditioners. Our work presents strategies to determine when multiprecision GMRES will be effective and to choose parameters for a multiprecision iterative refinement solver to achieve better performance. We use an implementation that is based on the Trilinos library and employs Kokkos Kernels for performance portability of linear algebra kernels. Performance results demonstrate the promise of multiprecision approaches and demonstrate even further improvements are possible by optimizing low-level kernels.
Halappanavar, Mahantesh H.; Acer, Seher A.; Boman, Erik G.; Buluc, Aydin B.; Ekanayate, Saliya E.; Feerdous, SM F.; Gawande, Nitin G.; Ghosh, Sayan G.; Khan, Arif K.; Minotoli, Marco M.; pothen, alex p.; Rajamanickam, Sivasankaran R.; Selvitopi, Oguz S.; Tallent, Nathan T.; Tumeo, Antonio T.
This presentation concludes in situ computation enables new approaches to linear algebra problems which can be both more effective and more efficient as compared to conventional digital systems. Preconditioning is well-suited to analog computation due to the tolerance for approximate solutions. When combined with prior work on in situ MVM for scientific computing, analog preconditioning can enable significant speedups for important linear algebra applications.
Over the past decade as Moore's Law has slowed, the need for new forms of computation that can provide sustainable performance improvements has risen. A new method, called in situ computing, has shown great potential to accelerate matrix vector multiplication (MVM), an important kernel for a diverse range of applications from neural networks to scientific computing. Existing in situ accelerators for scientific computing, however, have a significant limitation: These accelerators provide no acceleration for preconditioning-A key bottleneck in linear solvers and in scientific computing workflows. This paper enables in situ acceleration for state-of-The-Art linear solvers by demonstrating how to use a new in situ matrix inversion accelerator for analog preconditioning. As existing techniques that enable high precision and scalability for in situ MVM are inapplicable to in situ matrix inversion, new techniques to compensate for circuit non-idealities are proposed. Additionally, a new approach to bit slicing that enables splitting operands across multiple devices without external digital logic is proposed. For scalability, this paper demonstrates how in situ matrix inversion kernels can work in tandem with existing domain decomposition techniques to accelerate the solutions of arbitrarily large linear systems. The analog kernel can be directly integrated into existing preconditioning workflows, leveraging several well-optimized numerical linear algebra tools to improve the behavior of the circuit. The result is an analog preconditioner that is more effective (up to 50% fewer iterations) than the widely used incomplete LU factorization preconditioner, ILU(0), while also reducing the energy and execution time of each approximate solve operation by 1025x and 105x respectively.
Parallel implementations of linear iterative solvers generally alternate between phases of data exchange and phases of local computation. Increasingly large problem sizes and more heterogeneous compute architectures make load balancing and the design of low latency network interconnects that are able to satisfy the communication requirements of linear solvers very challenging tasks. In particular, global communication patterns such as inner products become increasingly limiting at scale. We explore the use of asynchronous communication based on one-sided Message Passing Interface primitives in the context of domain decomposition solvers. In particular, a scalable asynchronous two-level Schwarz method is presented. We discuss practical issues encountered in the development of a scalable solver and show experimental results obtained on a state-of-the-art supercomputer system that illustrate the benefits of asynchronous solvers in load balanced as well as load imbalanced scenarios. Using the novel method, we can observe speedups of up to four times over its classical synchronous equivalent.