# Tag Info

17

For what it is worth, for random sparse matrices of size 10,000 by 10,000 vs. dense matrices of the same size, on my Xeon workstation using MATLAB and Intel MKL as the BLAS, the sparse matrix-vector multiply was faster for densities of 15% or less. At 67% (as proposed by another answer), the dense matrix-vector multiplication was about three time faster.

14

It's a question on what you spend your time on. For most of us, we spend 3/4 of the time programming and 1/4 of the time waiting for results. (Your numbers may vary, but I think the number is not completely without merits.) So, if you have a design that allowed you to program twice as fast (3/4 of a time unit instead of 1.5 time units), then you can can take ...

13

When you use ZGELSS to sovle this problem, you're using the truncated singular value decomposition to regularize this extremely ill-conditioned problem. it's important to understand that this library routine is not attempting to find a least squares solution to $Ax=b$, but rather it is attempting to balance finding a solution that minimizes $\| x \|$ ...

13

BLAS1-operations, BLAS2-operations, and sparse-operations share the same curse of low arithmetic intensity, that they perform $O(1)$ flops for each memory read (contrast this to a BLAS3-operation like gemm, which performs $O(N^3)$ flops over $O(N^2)$ reads and only becomes more and more arithmetic-intensive/compute-bound for large $N$). However, sparse ...

12

The degeneracy of some eigenvalues looks to me like the hallmark of the breakdown of the Lanczos algorithm. The Lanczos algorithm is one of the more commonly used methods to approximate the eigenvalues and eigenvectors of Hermitian matrices; it's what scipy.eigsh() uses, through a call to the ARPACK library. In exact arithmetic, the Lanczos algorithm ...

12

The LU factors of a sparse matrix are at least somewhat sparse. The $Q$ matrix in QR can also somewhat preserve sparsity, and is typically used when the matrix is very long and skinny. The SVD of a sparse matrix will almost always have fully dense $U$ and $V$ factors, so it destroys any reason to perform the computations treating the matrix sparsely.

12

I believe comparing an iterative method (multigrid) to a direct/exact method (Thomas) in terms of exact operation count isn't really meaningful. IIRC, Thomas operation count is $8N$ for any tridiagonal system. The only time I can imagine multigrid conceivably beating that is for a trivial case of having a linear solution, and even then the cost of evaluating ...

12

Defining the auxiliary variable $y=Bx$ yields the following algebraically equivalent expanded system, $$\underbrace{\begin{bmatrix} 0 & A \\ B & -I \end{bmatrix}}_{K} \underbrace{\begin{bmatrix} x \\ y \end{bmatrix}}_{u} = \underbrace{\begin{bmatrix} b \\ 0 \end{bmatrix}}_{f},$$ which you could solve with GMRES or another nonsymmetric Krylov method. ...

11

For eigenvalues, simply take $k$ largest or smallest eigenvalues of $T$. They are good approximations of $A$, provided that the number of Lanczos iterations is large compared to $k$. Things are a little trickier if we want eigenvectors as well. The simplest way is to multiply each eigenvector $\mathbf{u}_i$ of $T$ by $V$ to the left, where $V$ is, as you ...

11

The short answer is that the Thomas algorithm will be faster than any iterative scheme for almost all cases. The exception would perhaps be applying a single iteration of a very simple iterative scheme such as Gauss-Seidel, but this is highly unlikely to give an acceptable solution. Also, this is ignoring parallel processing concerns. Multigrid is an ...

11

In deal.II (http://www.dealii.org -- disclaimer: I'm one of the principal authors of that library), we do eliminate whole rows and columns, and it is not too expensive overall. The trick is to use the fact that the sparsity pattern is typically symmetric, so you know which rows you need to look into when eliminating a whole column. The better approach, in ...

11

All matrix operations are memory bound (and not compute bound) on today's processors. So basically, you have to ask which format stores fewer bytes. This is easy to compute: For a full matrix, you store 8 bytes (one double) per entry For a sparse matrix, you store 12 bytes per entry (one double for the value, and one integer for the column index of the ...

10

Without taking sides the discussion about whether to use direct or iterative solvers, I just want to add two points: There exist Krylov methods for systems with multiple right-hand sides (called block Krylov methods). As an added bonus, these often have faster convergence than standard Krylov methods since the Krylov space is built from a larger collection ...

10

This is called "structurally symmetric". It simplifies graph traversal, such as occurs when setting up aggregates in algebraic multigrid, but doesn't offer much structure to improve convergence rates. Note that all common PDE discretizations have this property so this is still a huge class of matrices including many instances for which no truly good ...

10

Iterative Krylov-subspace solvers generally only require matrix-vector products and don't care whether or where there are zeros in the matrix. In your case, unless you have other information about the matrix (e.g., symmetry), you could for example use GMRES. What you probably had in mind is the question of preconditioning, and that you can't use things such ...

9

There is typically a trade-off between the amount of work you put into constructing a good preconditioner for an iterative solver and the work you save by using a good preconditioner when actually solving the linear systems. In your case, the case is pretty clear: put as much work as you can into constructing a good preconditioner because you have to solve ...

9

The MUMPS sparse direct solver can handle symmetric indefinite systems and is freely available (http://graal.ens-lyon.fr/MUMPS/). Ian Duff was one of the authors of both MUMPS and MA57 so the algorithms have many similarities. MUMPS was designed for distributed-memory parallel computers but it also works well on single-processor machines. If you link it ...

9

Jed Brown has already pointed this out in the comments to the question, but there is really not very much you can do in usual double precision if your condition number is large: in most cases, you will likely not get a single digit of accuracy in your solution and, worse, you can't even tell because you can't accurately evaluate the residual corresponding to ...

9

Very few scientific software developers understand good principles of design, so I apologize if this answer is a bit long-winded. From a software engineering perspective, the goal of the scientific software developer is to design a solution that satisfies a set of constraints that are often conflicting. Here are some typical examples of these constraints, ...

9

You can use the shift-invert spectral transform [1] and compute the spectrum band by band. The technique is also explained in my article [2]. Besides the implementation in [1], an implementation is available in C++ in my Graphite software [3] (update Jan 17: now everything is ported to geogram/graphite version 3), that I used to compute the eigenfunctions ...

9

Eigen 3 is a nice C++ template library some of whose routines are parallelized. c.f. Eigen documentation The parallelization is OMP only, so if you intend to parallelise using MPI (and OMP) it is probably not suitable for your purpose. The nice feature of Eigen is that you can swap in a high performance BLAS library (like MKL or OpenBLAS) for some routines ...

9

An anti-Hermitian matrix is diagonalizable, with orthogonal eigenvectors (ref). Hence you can write $X = PDP^{-1}$, where $D$ is a diagonal matrix. Therefore the exponential can be calculated as $e^X=Pe^DP^{-1}$, and $e^{tX} = Pe^{tD}P^{-1}$. If $d_1$, $d_2$, etc.... are the diagonal elements of $D$, then for each value of $t_i \in t$, $e^{t_iD}$ is just a ...

8

[The comment thread is long enough to convert to an answer.] We acknowledge that most matrix collections don't include physically reasonable right hand sides. If we want to test solver performance, we have to generate problems to solve --- either the right-hand sides $b$ or solutions $x$ from which we compute $b \gets A x$ and then solve $A y = b$ to some ...

8

The matrix product $B = AA^T$ is generally faster to apply as $A (A^T x)$ even when the product matrix $B$ is already available. Only for peculiar graphs with extremely low expansion factor would the product matrix be faster. PDE graphs in two or more dimensions have higher expansion factors. For example, if $A$ is a 9-point differencing operation on a 2D ...

8

Yes, the most common approach is to rebuild. Data structures that are modifiable in-place tend to be less efficient once set up, and reallocation is actually quite cheap compared to reassembly (e.g., due to nonlinearity) so it's really a fine solution. Outside of relatively rare niches with very easy solves, attempts to use dynamic data structures in the ...

8

The simple answer is that you would use inverse iteration (subspace or with deflation). This is basically the power method (repeatedly multiplying the matrix by a vector and normalizing, singling out the eigenvector corresponding to the eigenvalue of largest magnitude) applied to $A^{-1}$. Since you desire the $k$ closest to the origin, you need to use some ...

8

When looking at the solution of your system, you will find that almost all entries of $x$ are nonzero although the right-hand side is "sparse". Hence, whatever algorithm you use, it'll have to visit each and every entry at least once, so one wouldn't expect that you can save a lot of time using the sparsity of $b$. Right-hand sides where you can save a lot ...

8

There are two relatively convenient options for calculating selected (e.g. a few largest or smallest) eigenvalues using Eigen. The first is Spectra, a header-only C++ library based on Eigen that uses algorithms similar to ARPACK (implicitly-restarted Arnoldi) to calculate a few eigensolutions. Since it is header-only, you simply download and include the ...

8

There is a nice discussion on StackOverflow regarding floating point vs integer operations. In short, the performance of the operations depends a lot on processor architecture how the data is stored in memory and in which order it is accessed if (and which) SSE/AVX/AVX2/etc instructions are used (and how efficiently) This probably provides some insight ...

8

In theory, as the original authors, you're free to pick and name a standard, then expect others to follow it. In practise, if you're supporting an HPC system, then your choice is likely to be restricted among the compiler standards that the given system you're using for testing has for its tool stack (you are testing your software, aren't you?). This might ...

Only top voted, non community-wiki answers of a minimum length are eligible