arXiv subjects

Jonas Thies

Publications and source records attributed to Jonas Thies.

13 recordsLinked to original sources

arXiv2026

Implementation of QR factorization of tall and very skinny matrices on current GPUs

We consider the problem of computing a QR (or QZ) decomposition of a real, dense, tall and very skinny matrix. That is, the number of columns is tiny compared to the number of rows, rendering most computations completely or partially memory-bandwidth limited. The paper focuses on recent NVIDIA GPGPUs still supporting 64-bit floating-point arithmetic, but the findings carry over to AMD GPUs as well. We discuss two basic algorithms: Methods based on the normal equations (Gram matrix), in particular Cholesky-QR2 and SVQB, and the "tall-skinny QR" (TSQR), based on Householder transformations in a tree-reduction scheme. We propose two primary optimization techniques: Avoiding the write-back of the Q factor ("Q-less QR"), and exploiting fast local memory (shared memory on GPUs). We compare a straight-forward implementation of Gramian-based methods, and a more sophisticated TSQR implementation, in terms of performance achieved, time-to-solution, and implementation complexity. By performance modelling and numerical experiments with our own code and a vendor-optimized library routine, we demonstrate the crucial need for specialized methods and implementations in this memory-bound to transitional (memory/compute-bound) regime, and that TSQR is competitive in terms of time-to-solution, but at the cost of an investment in low-level code optimization.

cs.MS↗

arXiv2023

Performance of linear solvers in tensor-train format on current multicore architectures

Tensor networks are a class of algorithms aimed at reducing the computational complexity of high-dimensional problems. They are used in an increasing number of applications, from quantum simulations to machine learning. Exploiting data parallelism in these algorithms is key to using modern hardware. However, there are several ways to map required tensor operations onto linear algebra routines ("building blocks"). Optimizing this mapping impacts the numerical behavior, so computational and numerical aspects must be considered hand-in-hand. In this paper we discuss the performance of solvers for low-rank linear systems in the tensor-train format (also known as matrix-product states). We consider three popular algorithms: TT-GMRES, MALS, and AMEn. We illustrate their computational complexity based on the example of discretizing a simple high-dimensional PDE in, e.g., $50^{10}$ grid points. This shows that the projection to smaller sub-problems for MALS and AMEn reduces the number of floating-point operations by orders of magnitude. We suggest optimizations regarding orthogonalization steps, singular value decompositions, and tensor contractions. In addition, we propose a generic preconditioner based on a TT-rank-1 approximation of the linear operator. Overall, we obtain roughly a 5x speedup over the reference algorithm for the fastest method (AMEn) on a current multicore CPU.

math.NA↗

arXiv2023

Algebraic Temporal Blocking for Sparse Iterative Solvers on Multi-Core CPUs

Sparse linear iterative solvers are essential for many large-scale simulations. Much of the runtime of these solvers is often spent in the implicit evaluation of matrix polynomials via a sequence of sparse matrix-vector products. A variety of approaches has been proposed to make these polynomial evaluations explicit (i.e., fix the coefficients), e.g., polynomial preconditioners or s-step Krylov methods. Furthermore, it is nowadays a popular practice to approximate triangular solves by a matrix polynomial to increase parallelism. Such algorithms allow to evaluate the polynomial using a so-called matrix power kernel (MPK), which computes the product between a power of a sparse matrix A and a dense vector x, or a related operation. Recently we have shown that using the level-based formulation of sparse matrix-vector multiplications in the Recursive Algebraic Coloring Engine (RACE) framework we can perform temporal cache blocking of MPK to increase its performance. In this work, we demonstrate the application of this cache-blocking optimization in sparse iterative solvers. By integrating the RACE library into the Trilinos framework, we demonstrate the speedups achieved in preconditioned) s-step GMRES, polynomial preconditioners, and algebraic multigrid (AMG). For MPK-dominated algorithms we achieve speedups of up to 3x on modern multi-core compute nodes. For algorithms with moderate contributions from subspace orthogonalization, the gain reduces significantly, which is often caused by the insufficient quality of the orthogonalization routines. Finally, we showcase the application of RACE-accelerated solvers in a real-world wind turbine simulation (Nalu-Wind) and highlight the new opportunities and perspectives opened up by RACE as a cache-blocking technique for MPK-enabled sparse solvers.

math.NA↗

arXiv2022

Tensor Product Scheme for Computing Bound States of the Quantum Mechanical Three-Body Problem

We develop a computationally and numerically efficient method to calculate binding energies and corresponding wave functions of quantum mechanical three-body problems in low dimensions. Our approach exploits the tensor structure of the multidimensional stationary Schrödinger equation, being expressed as a discretized linear eigenvalue problem. In one spatial dimension, we solve the three-body problem with the help of iterative methods. Here the application of the Hamiltonian operator is represented by dense matrix-matrix products. In combination with a newly-designed preconditioner for the Jacobi-Davidson QR, our highly accurate tensor method offers a significantly faster computation of three-body energies and bound states than other existing approaches. For the two-dimensional case, we additionally make use of a hybrid distributed/shared memory parallel implementation to calculate the corresponding three-body energies. Our novel method is of high relevance for the analysis of few-body systems and their universal behavior, which is only governed by the particle masses, overall symmetries, and the spatial dimensionality. Our results have straightforward applications for ultracold atomic gases that are widespread and nowadays utilized in quantum sensors.

physics.comp-ph↗

arXiv2022

Performance of the low-rank tensor-train SVD (TT-SVD) for large dense tensors on modern multi-core CPUs

There are several factorizations of multi-dimensional tensors into lower-dimensional components, known as `tensor networks'. We consider the popular `tensor-train' (TT) format and ask: How efficiently can we compute a low-rank approximation from a full tensor on current multi-core CPUs? Compared to sparse and dense linear algebra, kernel libraries for multi-linear algebra are rare and typically not as well optimized. Linear algebra libraries like BLAS and LAPACK may provide the required operations in principle, but often at the cost of additional data movements for rearranging memory layouts. Furthermore, these libraries are typically optimized for the compute-bound case (e.g.\ square matrix operations) whereas low-rank tensor decompositions lead to memory bandwidth limited operations. We propose a `tensor-train singular value decomposition' (TT-SVD) algorithm based on two building blocks: a `Q-less tall-skinny QR' factorization, and a fused tall-skinny matrix-matrix multiplication and reshape operation. We analyze the performance of the resulting TT-SVD algorithm using the Roofline performance model. In addition, we present performance results for different algorithmic variants for shared-memory as well as distributed-memory architectures. Our experiments show that commonly used TT-SVD implementations suffer severe performance penalties. We conclude that a dedicated library for tensor factorization kernels would benefit the community: Computing a low-rank approximation can be as cheap as reading the data twice from main memory. As a consequence, an implementation that achieves realistic performance will move the limit at which one has to resort to randomized methods that only process part of the data.

math.NA↗

arXiv2021

(R)SE challenges in HPC

We discuss some specific software engineering challenges in the field of high-performance computing, and argue that the slow adoption of SE tools and techniques is at least in part caused by the fact that these do not address the HPC challenges `out-of-the-box'. By giving some examples of solutions for designing, testing and benchmarking HPC software, we intend to bring software engineering and HPC closer together.

cs.SE↗

arXiv2020

A staggered-grid multilevel incomplete LU for steady incompressible flows

Algorithms for studying transitions and instabilities in incompressible flows typically require the solution of linear systems with the full Jacobian matrix. Other popular approaches, like gradient-based design optimization and fully implicit time integration, also require very robust solvers for this type of linear system. We present a parallel fully coupled multilevel incomplete factorization preconditioner for the 3D stationary incompressible Navier-Stokes equations on a structured grid. The algorithm and software are based on the robust two-level method developed by Wubs and Thies. In this paper, we identify some of the weak spots of the two-level scheme and propose remedies such as a different domain partitioning and recursive application of the method. We apply the method to the well-known 3D lid-driven cavity benchmark problem, and demonstrate its superior robustness by comparing with a segregated SIMPLE-type preconditioner.

math.NA↗

arXiv2020

Performance Engineering for Real and Complex Tall & Skinny Matrix Multiplication Kernels on GPUs

General matrix-matrix multiplications with double-precision real and complex entries (DGEMM and ZGEMM) in vendor-supplied BLAS libraries are best optimized for square matrices but often show bad performance for tall & skinny matrices, which are much taller than wide. NVIDIA's current CUBLAS implementation delivers only a fraction of the potential performance as indicated by the roofline model in this case. We describe the challenges and key characteristics of an implementation that can achieve close to optimal performance. We further evaluate different strategies of parallelization and thread distribution, and devise a flexible, configurable mapping scheme. To ensure flexibility and allow for highly tailored implementations we use code generation combined with autotuning. For a large range of matrix sizes in the domain of interest we achieve at least 2/3 of the roofline performance and often substantially outperform state-of-the art CUBLAS results on an NVIDIA Volta GPGPU.

cs.MS↗

arXiv2019

A Recursive Algebraic Coloring Technique for Hardware-Efficient Symmetric Sparse Matrix-Vector Multiplication

The symmetric sparse matrix-vector multiplication (SymmSpMV) is an important building block for many numerical linear algebra kernel operations or graph traversal applications. Parallelizing SymmSpMV on today's multicore platforms with up to 100 cores is difficult due to the need to manage conflicting updates on the result vector. Coloring approaches can be used to solve this problem without data duplication, but existing coloring algorithms do not take load balancing and deep memory hierarchies into account, hampering scalability and full-chip performance. In this work, we propose the recursive algebraic coloring engine (RACE), a novel coloring algorithm and open-source library implementation, which eliminates the shortcomings of previous coloring methods in terms of hardware efficiency and parallelization overhead. We describe the level construction, distance-k coloring, and load balancing steps in RACE, use it to parallelize SymmSpMV, and compare its performance on 31 sparse matrices with other state-of-the-art coloring techniques and Intel MKL on two modern multicore processors. RACE outperforms all other approaches substantially and behaves in accordance with the Roofline model. Outliers are discussed and analyzed in detail. While we focus on SymmSpMV in this paper, our algorithm and software is applicable to any sparse matrix operation with data dependencies that can be resolved by distance-k coloring.

cs.DC↗

arXiv2018

Benefits from using mixed precision computations in the ELPA-AEO and ESSEX-II eigensolver projects

We first briefly report on the status and recent achievements of the ELPA-AEO (Eigenvalue Solvers for Petaflop Applications - Algorithmic Extensions and Optimizations) and ESSEX II (Equipping Sparse Solvers for Exascale) projects. In both collaboratory efforts, scientists from the application areas, mathematicians, and computer scientists work together to develop and make available efficient highly parallel methods for the solution of eigenvalue problems. Then we focus on a topic addressed in both projects, the use of mixed precision computations to enhance efficiency. We give a more detailed description of our approaches for benefiting from either lower or higher precision in three selected contexts and of the results thus obtained.

physics.comp-ph↗

arXiv2017

CRAFT: A library for easier application-level Checkpoint/Restart and Automatic Fault Tolerance

In order to efficiently use the future generations of supercomputers, fault tolerance and power consumption are two of the prime challenges anticipated by the High Performance Computing (HPC) community. Checkpoint/Restart (CR) has been and still is the most widely used technique to deal with hard failures. Application-level CR is the most effective CR technique in terms of overhead efficiency but it takes a lot of implementation effort. This work presents the implementation of our C++ based library CRAFT (Checkpoint-Restart and Automatic Fault Tolerance), which serves two purposes. First, it provides an extendable library that significantly eases the implementation of application-level checkpointing. The most basic and frequently used checkpoint data types are already part of CRAFT and can be directly used out of the box. The library can be easily extended to add more data types. As means of overhead reduction, the library offers a build-in asynchronous checkpointing mechanism and also supports the Scalable Checkpoint/Restart (SCR) library for node level checkpointing. Second, CRAFT provides an easier interface for User-Level Failure Mitigation (ULFM) based dynamic process recovery, which significantly reduces the complexity and effort of failure detection and communication recovery mechanism. By utilizing both functionalities together, applications can write application-level checkpoints and recover dynamically from process failures with very limited programming effort. This work presents the design and use of our library in detail. The associated overheads are thoroughly analyzed using several benchmarks.

cs.DC↗

arXiv2016

GHOST: Building blocks for high performance sparse linear algebra on heterogeneous systems

While many of the architectural details of future exascale-class high performance computer systems are still a matter of intense research, there appears to be a general consensus that they will be strongly heterogeneous, featuring "standard" as well as "accelerated" resources. Today, such resources are available as multicore processors, graphics processing units (GPUs), and other accelerators such as the Intel Xeon Phi. Any software infrastructure that claims usefulness for such environments must be able to meet their inherent challenges: massive multi-level parallelism, topology, asynchronicity, and abstraction. The "General, Hybrid, and Optimized Sparse Toolkit" (GHOST) is a collection of building blocks that targets algorithms dealing with sparse matrix representations on current and future large-scale systems. It implements the "MPI+X" paradigm, has a pure C interface, and provides hybrid-parallel numerical kernels, intelligent resource management, and truly heterogeneous parallelism for multicore CPUs, Nvidia GPUs, and the Intel Xeon Phi. We describe the details of its design with respect to the challenges posed by modern heterogeneous supercomputers and recent algorithmic developments. Implementation details which are indispensable for achieving high efficiency are pointed out and their necessity is justified by performance measurements or predictions based on performance models. The library code and several applications are available as open source. We also provide instructions on how to make use of GHOST in existing software packages, together with a case study which demonstrates the applicability and performance of GHOST as a component within a larger software stack.

cs.DC↗

arXiv2010

A robust two-level incomplete factorization for (Navier-)Stokes saddle point matrices

We present a new hybrid direct/iterative approach to the solution of a special class of saddle point matrices arising from the discretization of the steady incompressible Navier-Stokes equations on an Arakawa C-grid. The two-level method introduced here has the following properties: (i) it is very robust, even close to the point where the solution becomes unstable; (ii) a single parameter controls fill and convergence, making the method straightforward to use; (iii) the convergence rate is independent of the number of unknowns; (iv) it can be implemented on distributed memory machines in a natural way; (v) the matrix on the second level has the same structure and numerical properties as the original problem, so the method can be applied recursively; (vi) the iteration takes place in the divergence- free space, so the method qualifies as a 'constraint preconditioner'; (vii) the approach can also be applied to Poisson problems. This work is also relevant for problems in which similar saddle point matrices occur, for instance when simulating electrical networks, where one has to satisfy Kirchhoff's conservation law for currents.

math.NA↗