I. S. Duff, R. G. Grimes, and J. G. Lewis, User's guide for the Harwell-Boeing sparse matrix collection (release I), Tech. Report TR/PA/92/86, CERFACS, Toulouse Cedex, France, Oct. 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Receive(A i s i ; p i ) fs i computed using r at t 1 g r(t 1) r(t) i A i s i : Check for termination. if not termination then t = t 1; Send(r(t) p i ) ft 2 = tg 4 Experiments Test problems used in the experiments are taken from Harwell Boeing sparse matrix test collection [5]. All the graphs refer to problem ASH958. Blocks are formed by taking blocks of consecutive columns. For the specified problem, the number of blocks g is 30. Each subproblem is solved using QR factorization and SNE. Both parallel and sequential implementations are done on Intel Paragon. Static ....

Duff, I., Grimes, R. G., Lewis, J. G.: User's Guide for the Harwell-Boeing Sparse Matrix Collection. Technical Report TR/PA/92/86, CERFACS (1992)

....coming from a finite element structural analysis; a SVD estimation for a rectangular matrix coming from an economics analysis model; and a problem related with the investigation of motions of a protein. We also include a set of cases available in the Hatwell Boeing Sparse Matrix Collection [12], to show the effects of the block size and the spectrum slicing strategy. We use the notation of the previous sections: ra denotes the number of required solutions, n the dimension of the matrices and p the block size for the Lanczos algorithm. Except for the third application, all computations ....

....(55 55) 27 30) 23 30) 19 31) 16 34) 14 34) 16 4.2 A SVD estimation In this application we deal with a standard eigenproblem. In particular, BLZPACK is employed to estimate some singular values and vectors of the matrix WM3, which is available in the Hatwell Boeing Sparse Matrix Collection [12]. The singular value decomposition is a clever way of determining the matrix rank and it is therefore a useful tool in the analysis of systems of equations [15] WM3 is related with a model of the world economy and has dimension 207 x 260. Its pattern is given in Figure 10, which contains 2948 ....

[Article contains additional citation context not shown here]

I. S. Duff, R. G. Grimes, and J. G. Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection (Release I). Technical Report TR/PA/92/86, CERFACS, Toulouse, France, 1992.

....in the previous stage and ordering the nodes in each subtree consecutively [11] We will call this algorithm LNS (left looking algorithm using the new scheduling) 5 Results We have implemented the codes on the SGI 02000 using parmacs. A set of symmetric matrices from the Harwell Boeing library [2] was selected to analyze the behaviour of the codes. The minimum degree algorithm was previously applied in order to reduce the fill. Table 1 shows the characteristics of the matrices, where nz is the number of nonzero entries, and n is the number of columns. MATRIX n nz in A nz in L BCSSTK15 ....

I.S.Duff, R.G. Grimes, and J.G.Lewis, User's guide for the Harwell-Boeing sparse matrix collection, Technical Report TR-PA-92-96, CERFACS, October 1992.

....process simulation, finite element and finite difference applications, optimization problems. Therefore, we have used in our experiments as indirection arrays the patterns of sparse matrices that appear in real codes. These matrices were extracted from the HarwellBoeing (HB) collection [4]. We have also considered synthetic access patterns to cover a wider range of cases. The patterns (5 HB and 4 synthetic) are characterized in Table 1, where 2 N is the size of the indirection array INDEX and M is the size of array A. The parameter CP (Critical Path, 1 CP N ) is the length of ....

I. S. Duff, R. G. Grimes, and J. G. Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection. Technical Report TR-PA-92-96, CERFACS, 1992.

....ideas of this paper can also be applied to alternating direction line SOR methods with little or no modification. Alternating direction line SOR methods are often used as smoothers for multigrid methods since their smoothing properties are slightly better than the corresponding pointwise smoothers [5] (a smoothing rate of 0:2 for line SOR vs a smoothing rate of 0:25 for point SOR) ....

I. S. Duff, R. G. Grimes, and J. G. Lewis. User's guide for the Harwell--Boeing sparse matrix collection (Release I). Technical Report TR/PA/92/86, CERFACS, Toulouse, 1992.

....can be returned to the system for use in subsequent stages (which would use more memory than even the dynamically allocated graph at any rate) 6 Experimental Results 6. 1 Test Matrices We tested the performance of our implementation using selected matrices from the Harwell Boeing collection [16], the University of Florida s sparse matrix collection [17] as well as locally generated matrices representing discretized Laplacians. For our tests, we compare the execution time of our implementation against that of the equivalent SPARSPAK algorithm (GENMMD) The tests were run on a Sun SPARC ....

Roger G. Grimes, John G. Lewis, and Iain S. Duff. User's guide for the harwell-boeing sparse matrix collection. User's Manual Release 1, Boeing Computer Services, Seattle, WA, October 1992.

....(PDE225) More information about the PDEN matrices can be found in 79 the user s guide for the NEP matrix collection [3] which is available from the Matrix Market web site. Other matrices. The other matrices used in numerical experiments are FS1836 from the Harwell Boeing matrix collection [13] and BWM200 1 from the NEP matrix collection. Both are real, nonsymmetric and were downloaded from the Matrix Market web site. Matrix statistics: A N (A) nnz(A) XA ) FS1836 183 1.74e 11 1000 5.62e 17 BWM200 200 3.39 796 2.23 B MATLAB code for Phase II of the ABF method In Appendix B we ....

I. S. Duff, R. G. Grimes, and J. G. Lewis, User's guide for the Harwell-Boeing sparse matrix collection (Release I), Tech. Report TR/PA/92/86, CERFACS, 1992.

....fast a Ethernet network. We developed our programs using MATLAB and conducted our experiments over a novel environment that allows the concurrent operation of MATLAB; see the next subsection. We use standard test matrices, drawn from the Test Matrix Toolbox ( 9] and the Harwell Boeing collection ([7]) throughout our discussion. 1.1.1 Programming environment We used the Cornell Multitask Toolbox [19] developed by J.A. Zollweg and A. Verma at the Cornell Theory Center, for MATLAB (version 5.3) the widely used problem solving environment from Mathworks. The multitasking toolbox enables ....

I.S. Duff, R.G. Grimes, and J.G. Lewis. User's guide for the Harwell-Boeing sparse matrix collection (Release I). Tech. Rep. TR/PA/92/86, CERFACS, Toulouse Cedex, France, Oct. 1992.

.... 1e 06 1e 04 1e 02 1 1e02 0 100 200 300 400 500 600 700 800 900 iterations FOM GMRES WFOM WGMRES Figure 3: The matrix add20, with m=10 Example 2: the matrix orsirr 1 It is a 1030 x 1030 matrix arising from a 3D oil reservoir simulation, from the package OILGEN of the Harwell Boeing collection [5], with 6858 nonzero entries. The estimated condition number is 10e 2. In the Table 1, we report the results given for different values of m. We take epsilon = 10 Gamma11 and the algorithm is stopped if the convergence is not reached after 2000 iterations. We remark that, for all considered ....

I. S. Duff, R. G. Grimes, J. G. Lewis, User's guide for the HarwellBoeing sparse matrix collection (Release I), Tech. Rep. TR/PA/92/86, CERFACS, Toulouse, France, 1992.

....in order to avoid an excessive interprocessor traffic computation ratio. 5 Evaluation We implemented the parallel algorithm on the Fujitsu AP1000 and the CRAY T3D and T3E multicomputers, measuring the execution time in meshes of different sizes, using matrices from the Harwell Boeing collection [5, 6]. We also implemented the parallel algorithm using the message passing library of the AP1000 (CellOS) and the MPI standard library. In the results we present, we employed three matrices: BCSPWR10, BCSSTK18 and BCSSTK21 (Table 1) Even though these matrices are symmetric, they have been stored ....

Iain S. Duff, Roger G. Grimes, and John G. Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection. Research and Technology Division, Boeing Computer Services, October 1992.

....in order to avoid an excessive interprocessor traffic computation ratio. 3 Evaluation We implemented the parallel algorithm on the Fujitsu AP1000 and the CRAY T3D and T3E multicomputers, measuring the execution time in meshes of different sizes, using matrices from the Harwell Boeing collection [6,7]. We implemented the parallel algorithm using the MPI standard library. In the results we present, three matrices have been employed: BCSPWR10, BCSSTK18 and BCSSTK21 (Table 1) Even though these matrices are symmetric, they have been stored whole, since our objective is to find a general ....

Iain S. Duff, Roger G. Grimes, and John G. Lewis. User's Guide for the HarwellBoeing Sparse Matrix Collection. Research and Technology Division, Boeing Computer Services, October 1992.

....distibuted according to a DSC distribution. Howeever, this scheme does not follow the DSC ordering strictly. When the next row in the schedule is not ready, we allow other rows to computed if they are ready for execution. 4. 5 Benchmarks We chose four power grid matrices from the Harwell Boeing [DGL92] benchmark set for our experiments. These sparse matrices come from the BCSPWR set, and represent parts of the Western US Power Network and the Eastern US Power Network. Our matrices were incompletely factored, with the sparsity pattern being identical to that of the original matrix. In Table 1, ....

Ian S. Duff, Roger G. Grimes, and John G. Lewis. User's guide for the Harwell-Boeing sparse matrix collection. Technical Report TR/PA/92/86, CERFACS, 42 Ave G. Coriolis, 31057 Toulouse Cedex, France, October 1992.

....ncomp is the number of components in the EC graph. Each of these ncomp guard elements is a list of the interface variables lying in the corresponding component of the EC graph. The improvements this can give may be illustrated using problem LOCK1074 from the Harwell Boeing sparse matrix collection (Duff, Grimes, and Lewis 1989, 1992). We partitioned the finite element domain into 4 using the code Chaco (see Section 5 for details) and generated the EC graph for each subdomain. The EC(g) graphs for subdomains 3 and 4 were found to have 3 and 2 components, respectively. Applying 12 Approach II applied to the single component ....

....assess the quality of the element reordering algorithms discussed in this paper. The problems range in size from 360 to 23446 elements. A brief description of each test problem is given in Table 5.3. Problems with origin HB were taken from the widely used Harwell Boeing sparse matrix collection (Duff, Grimes, and Lewis 1989, 1992). Those with origin AEAT were supplied by Andrew Cliffe of AEA Technology, Harwell Laboratory; those with origin DNVR came from A. C. Damhaug of Det Norske Veritas Research, Norway; and those with origin RALPAR were supplied by R. F. Fowler of the Rutherford Appleton Laboratory. For the RALPAR ....

Duff, I. S., Grimes, R. G. and Lewis, J. G. (1992). User's guide for the Harwell-Boeing sparse matrix collection. Rutherford Appleton Laboratory Report RAL-92-086.

....For these experiments, GGCL is 5 to 7 times faster than LEDA. 5. 2 Comparison to Special Purpose Library In additionn, we demonstrate the performance of a GGCL based implementation of the multiple mininum degree algorithm [13] using selected matrices from the Harwell Boeing collection [9] and the University of Florida s sparse matrix collection [2] Our tests compare the execution time of our implementation against that of the equivalent SPARSPAK Fortran algorithm (GENMMD) 7] For each case, our implementation and GENMMD produced identical orderings. Note that the performance of ....

Grimes, R. G., Lewis, J. G., and Duff, I. S. User's guide for the harwell-boeing sparse matrix collection. User's Manual Release 1, Boeing Computer Services, Seattle, WA, October 1992.

....for computing the right most (or left most) eigenvalues (algorithm S2) For ease of reference, all the algorithms implemented in this study are summarised in Table 4.1. Most of the problems we have used for performance testing were taken from the the Harwell Boeing sparse matrix collection (Duff, Grimes, and Lewis 1992). This collection is now widely accepted and used in testing and evaluating sparse matrix algorithms. The matrices we selected from the collection are unsymmetric assembled matrices with the values of the nonzeros supplied, although some of them 19 Table 4.1. Summary of the algorithms used in ....

....acceleration A3 Chebychev preconditioned Arnoldi AB3 Block Chebychev preconditioned Arnoldi (namely PORES2 and PORES3) have a symmetric pattern. The problems and some of their characteristics are listed in Table 4.2. For more details of the origins of these problems the user is referred to Duff et al. 1992). The only test problems used which are not currently included in the Harwell Boeing set are GRT200, GRT400, and TOLSA. GRT200 and GRT400 were supplied by Garratt (1990, 1991) These matrices arise from the detection of Hopf bifurcation points and the main concern for these problems is not the ....

Duff, I. S., Grimes, R. G., and Lewis, J. G. (1992). User's guide for the Harwell-Boeing sparse matrix collection (Release 1). Technical Report RAL 92-086, Rutherford Appleton Laboratory.

No context found.

I. S. Duff, R. G. Grimes, and J. G. Lewis, User's guide for the Harwell-Boeing sparse matrix collection (release I), Tech. Report TR/PA/92/86, CERFACS, Toulouse Cedex, France, Oct. 1992.

No context found.

I. DUFF, R. GRIMES, AND J. LEWIS, User's guide for the Harwell-Boeing sparse matrix collection (release I), Tech. Rep. TR/PA/92/86, CERFACS, Toulouse Cedex, France, Oct. 1992.

No context found.

I.S. Duff, R.G. Grimes and J.G. Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection. Technical Report TR-PA92 -96. CERFACS, Oct 1992.

No context found.

I.S. Duff, R.G. Grimes, J.G. Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection. Technical Report TR-PA-92-96. CERFACS, Oct 1992.

No context found.

I.S.Duff, R.G.Grimes, and J.G.Lewis. User's guide for the harwell-boeing sparse matrix collection. Technical Report TR-PA-92-96, CERFACS, October 1992. 16--11

No context found.

I.S.Duff, R.G.Grimes, and J.G.Lewis. User's Guide for the Harwell-Boeing Sparse Matrix Collection. Technical Report TR-PA-92-96, CERFACS, October 1992.

No context found.

I.S. Duff, R.G. Grimes and J.G. Lewis, "User's Guide for the Harwell-Boeing Sparse Matrix Collection," CERFACS TR-PA-92-96, Oct. 1992.

No context found.

Duff, I.S., Grimes, R.G., Lewis, J.G.: User's Guide for the Harwell-Boeing Sparse Matrix Collection. Tech. Report TR-PA-92-96, CERFACS (1992).

No context found.

I. S. Duff, R. G. Grimes, and J. G. Lewis. User's guide for the Harwell--Boeing sparse matrix collection (Release I). Technical Report TR/PA/92/86, CERFACS, Toulouse, 1992.

No context found.

I. S. Duff, R. G. Grimes, and J. G. Lewis, "User's guide for the Harwell-Boeing sparse matrix collection," Technical Report TR/PA/92/86, CERFACS, 42 Ave G. Coriolis, 31057 Toulouse Cedex, France, October 1992.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC