May, 2008 * changed UFcollection to add links to Yifan Hu's graph pictures to each web page. May 7, 2008 * added new matrices, 1878 to 1890 (some obtained in Fall 2007) 1878 Lee/fem_filter 1879 Mancktelow/viscorocks 1880 Rudnyi/water_tank 1881 Rucci/Rucci1 1882 McRae/ecology1 1883 McRae/ecology2 1884 NYPA/Maragal_1 1885 NYPA/Maragal_2 1886 NYPA/Maragal_3 1887 NYPA/Maragal_4 1888 NYPA/Maragal_5 1889 NYPA/Maragal_6 1890 NYPA/Maragal_7 April 1, 2008 * UFget updated: added UFget('refresh') option. Better handling of failure when internet connection is not active. Sept 6, 2007 * added matrix 1877 1877 Rost/RFdevice Aug 30, 2007 * added matrices 1874 to 1876 1874 HVDC/hvdc1 1875 HVDC/hvdc2 1876 Zaoui/kkt_power Aug 21, 2007 * added matrices 1870 to 1873 1870 YCheng/psse0 1871 YCheng/psse1 1872 YCheng/psse2 1873 Dehghani/light_in_tissue Aug 8, 2007 * added matrices 1864 to 1869 1864 Muite/Chebyshev1 1865 Muite/Chebyshev2 1866 Muite/Chebyshev3 1867 Muite/Chebyshev4 1868 Quaglino/viscoplastic1 1869 Quaglino/viscoplastic2 July 31, 2007 * added matrix 1863 1863 Rajat/Raj1 July 25, 2007 * added note to F1 matrix (stiffness matrix from an engine crankshaft) * added matrices 1861 to 1862 1861 Koutsovasilis/F2 1862 Szczerba/Ill_Stokes June 20, 2007 * added matrices 1857 to 1860: 1857 Sinclair/3Dspectralwave2 1858 QLi/crashbasis 1859 QLi/majorbasis 1860 Springer/ESOC June 14, 2007 * many *.mat files updated, to ensure nzmax(A)-nnz(A) is at most 1. No change to MM and RB files. May 31, 2007 * added matrices 1841 to 1856: 1841 Gleich/wb-cs-stanford 1842 Gleich/wb-edu 1843 Gleich/wikipedia-20051105 1844 Gleich/wikipedia-20060925 1845 Gleich/wikipedia-20061104 1846 Gleich/wikipedia-20070206 1847 UTEP/Dubcova1 1848 UTEP/Dubcova2 1849 UTEP/Dubcova3 1850 BenElechi/BenElechi1 1851 Botonakis/FEM_3D_thermal1 1852 Botonakis/FEM_3D_thermal2 1853 Wissgott/parabolic_fem 1854 Watson/chem_master1 1855 Watson/Baumann 1856 Sinclair/3Dspectralwave Feb 1, 2007 * Problem notes, and "kind" modified for wathen100 and wathen120 matrices. These are from gallery('wathen',100,x) where x = 100 or 120, They form 100-by-x 2D meshes with randomly generated values. Jan 30, 2007 * UF_Index.isND added to the index. isND(i) is 1 if matrix i comes from a problem with 2D or 3D geometry, 0 otherwise. * added Problem.kind to all problems. It is now a required field in the Problem struct. The field has been added to all files in all three formats (MATLAB, Matrix Market, and Rutherford-Boeing). Jan 27, 2007 * updated the UF_Index. Corrected amd_lnz and amd_flops for 5 matrices (previously -1; that is, they had not yet been computed) 916 vanHeukelum/cage15 1252 GHS_psdef/audikw_1 1453 Oberwolfach/bone010 1455 Oberwolfach/boneS10 1586 Schenk_AFE/af_shell10 The amd_lnz and amd_flops items in the index are now complete. * UF_Index entry for Koutsovasilis/F1 revised: the matrix is symmetric positive definite. Dec 12, 2006 * cage15 Matrix Market file posted (2GB filesize problem resolved). Dec 4, 2006 * updated SuiteSparse posted, including UFcollection and RBio (software for managing the collection) Nov 30, 2006 * Added UFgrep to the UFget toolbox. Nov 29, 2006 * Modified UFget, so that it downloads the UF_Index.mat file if it doesn't exist or can't be loaded. Nov 25, 2006 * Matrix web pages recreated. Nov 21, 2006 * UF_Index.ncc added to index. For a square matrix, ncc is the number of strongly connected components of the graph (ignoring the diagonal). For a rectangular matrix, ncc is the number of connected components of the bipartite graph. * removed extraneous notes from Pajek/FA Nov 20, 2006 * All *.mat files cleaned up; extra entries other than "Problem" slipped in when the author, date, and ed fields were added on Nov 8. This glitch had no effect on the MM and RB versions of the collection. * Reconstructed UF_Index from scratch. Some statistics have changed because of minor changes in the ordering methods. Other statistics have changed because of corrections to the matrices themselves made recently. Removed has_b, has_guess, has_x, has_Zeros, is_lp, has_coord, zdiag from UF_Index; added RBtype and cholcand. RBtype is the Rutherford/Boeing type of the matrix. It consists of 3 letters: 1) r: real, p: pattern, i: integer, c: complex 2) r: rectangular, u: unsymmetric, s: symmetric, h: Hermitian, z: skew-symmetric 3) always "a" (assembled) UF_Index.cholcand(id) is 1 if the matrix is symmetric or Hermitian, and all diagonal entries are real and greater than zero (thus a candidate for Cholesky factorization). All code used to create the index will be posted shortly. Nov 16, 2006 * Added 30 linear programming problems from Hans Mittelmann (id's 1645 to 1674), http://plato.asu.edu/ftp/lptestset 1645 Mittelmann/nug08-3rd 1646 Mittelmann/pds-30 1647 Mittelmann/pds-40 1648 Mittelmann/pds-50 1649 Mittelmann/pds-60 1650 Mittelmann/pds-70 1651 Mittelmann/pds-80 1652 Mittelmann/pds-90 1653 Mittelmann/pds-100 1654 Mittelmann/rail507 not related to Oberwolfach rail* 1655 Mittelmann/rail516 1656 Mittelmann/rail582 1657 Mittelmann/rail2586 1658 Mittelmann/rail4284 1659 Mittelmann/sgpf5y6 1660 Mittelmann/stormG2_1000 1661 Mittelmann/watson_1 1662 Mittelmann/watson_2 1663 Mittelmann/cont11_l 1664 Mittelmann/cont1_l 1665 Mittelmann/fome11 1666 Mittelmann/fome12 1667 Mittelmann/fome13 1668 Mittelmann/fome20 1669 Mittelmann/fome21 1670 Mittelmann/neos 1671 Mittelmann/neos1 1672 Mittelmann/neos2 1673 Mittelmann/neos3 1674 Mittelmann/spal_004 * Corrected Problem.ed for LPnetlib problems ('D. Day' should be 'D. Gay') * Added 358 linear programming problems from C. Meszaros, http://www.sztaki.hu/~meszaros/public_ftp/lptestset, id's 1675 to 1840 1675 Meszaros/gams10am 1676 Meszaros/gams30am 1677 Meszaros/gams60am 1678 Meszaros/gas11 1679 Meszaros/aa01 1680 Meszaros/aa03 1681 Meszaros/aa3 1682 Meszaros/aa4 1683 Meszaros/aa5 1684 Meszaros/aa6 1685 Meszaros/air02 1686 Meszaros/air03 1687 Meszaros/air04 1688 Meszaros/air05 1689 Meszaros/air06 1690 Meszaros/aircraft 1691 Meszaros/bas1lp 1692 Meszaros/baxter 1693 Meszaros/car4 1694 Meszaros/cari 1695 Meszaros/ch 1696 Meszaros/co5 1697 Meszaros/co9 1698 Meszaros/complex 1699 Meszaros/cq5 1700 Meszaros/cq9 1701 Meszaros/cr42 1702 Meszaros/crew1 1703 Meszaros/dano3mip 1704 Meszaros/dbic1 1705 Meszaros/dbir1 1706 Meszaros/dbir2 1707 Meszaros/df2177 1708 Meszaros/e18 1709 Meszaros/ex3sta1 1710 Meszaros/farm 1711 Meszaros/gams10a 1712 Meszaros/gams30a 1713 Meszaros/ge 1714 Meszaros/iiasa 1715 Meszaros/jendrec1 1716 Meszaros/kl02 1717 Meszaros/l9 1718 Meszaros/lp22 1719 Meszaros/lpl2 1720 Meszaros/lpl3 1721 Meszaros/mod2 1722 Meszaros/model1 1723 Meszaros/model2 1724 Meszaros/model3 1725 Meszaros/model4 1726 Meszaros/model5 1727 Meszaros/model6 1728 Meszaros/model7 1729 Meszaros/model8 1730 Meszaros/model9 1731 Meszaros/model10 1732 Meszaros/nemsafm 1733 Meszaros/nemscem 1734 Meszaros/nemsemm1 1735 Meszaros/nemsemm2 1736 Meszaros/nemspmm1 1737 Meszaros/nemspmm2 1738 Meszaros/nemswrld 1739 Meszaros/nl 1740 Meszaros/nw14 1741 Meszaros/p0033 1742 Meszaros/p0040 1743 Meszaros/p010 1744 Meszaros/p0201 1745 Meszaros/p0282 1746 Meszaros/p0291 1747 Meszaros/p0548 1748 Meszaros/p05 1749 Meszaros/p2756 1750 Meszaros/p6000 1751 Meszaros/pcb1000 1752 Meszaros/pcb3000 1753 Meszaros/pf2177 1754 Meszaros/primagaz 1755 Meszaros/problem 1756 Meszaros/progas 1757 Meszaros/qiulp 1758 Meszaros/r05 1759 Meszaros/refine 1760 Meszaros/rlfddd 1761 Meszaros/rlfdual 1762 Meszaros/rlfprim 1763 Meszaros/rosen1 1764 Meszaros/rosen2 1765 Meszaros/rosen7 1766 Meszaros/rosen8 1767 Meszaros/rosen10 1768 Meszaros/route 1769 Meszaros/seymourl 1770 Meszaros/slptsk 1771 Meszaros/south31 1772 Meszaros/stat96v1 1773 Meszaros/stat96v2 1774 Meszaros/stat96v3 1775 Meszaros/stat96v4 1776 Meszaros/stat96v5 1777 Meszaros/t0331-4l 1778 Meszaros/testbig 1779 Meszaros/ulevimin 1780 Meszaros/us04 1781 Meszaros/world 1782 Meszaros/zed 1783 Meszaros/de063155 1784 Meszaros/de063157 1785 Meszaros/de080285 1786 Meszaros/gen1 1787 Meszaros/gen2 1788 Meszaros/gen4 1789 Meszaros/gen 1790 Meszaros/iprob 1791 Meszaros/l30 1792 Meszaros/stoch_aircraft 1793 Meszaros/cep1 1794 Meszaros/deter0 1795 Meszaros/deter1 1796 Meszaros/deter2 1797 Meszaros/deter3 1798 Meszaros/deter4 1799 Meszaros/deter5 1800 Meszaros/deter6 1801 Meszaros/deter7 1802 Meszaros/deter8 1803 Meszaros/fxm2-6 1804 Meszaros/fxm2-16 1805 Meszaros/fxm3_6 1806 Meszaros/fxm3_16 1807 Meszaros/fxm4_6 1808 Meszaros/pgp2 1809 Meszaros/stormg2-125 1810 Meszaros/stormg2-27 1811 Meszaros/stormg2-8 1812 Meszaros/degme 1813 Meszaros/karted 1814 Meszaros/tp-6 1815 Meszaros/ts-palko The following Problems are sequences of closely related LP's: 1816 Meszaros/delf 36: delf000 to delf036 (excl. 016) 1817 Meszaros/kleemin 6: kleemin3 to kleemin8 1818 Meszaros/large 37: large000 to large036 1819 Meszaros/nsct 2: nsct1 and nsct2 1820 Meszaros/nsic 2: nsic1 and nsic2 1821 Meszaros/nsir 2: nsir1 and nsir2 1822 Meszaros/plddb 13: pldd000b to pldd012b 1823 Meszaros/rat 3: rat1, rat5, and rat7a 1824 Meszaros/small 17: small000 to small017 1825 Meszaros/pltexpa 5: pltexpa2-6, 2-16, 3_6, 3_16, 4_6 1826 Meszaros/sc205-2r 12: sc205-2r-4 to 1600 1827 Meszaros/scagr7-2b 3: scagr7-2b-4 to 64 1828 Meszaros/scagr7-2c 3: scagr7-2c-4 to 64 1829 Meszaros/scagr7-2r 11: scagr7-2r-4 to 864 1830 Meszaros/scfxm1-2b 3: scfmx1-2b-4 to 64 1831 Meszaros/scfxm1-2r 9: scfmx1-2r-4 to 256 1832 Meszaros/scrs8-2b 3: scrs8-2b-4 to 64 1833 Meszaros/scrs8-2c 5: scrs8-2c-4 to 64 1834 Meszaros/scrs8-2r 10: scrs8-2r-4 to 512 1835 Meszaros/scsd8-2b 3: scsd8-2b-4 to 64 1836 Meszaros/scsd8-2c 3: scsd8-2c-4 to 64 1837 Meszaros/scsd8-2r 11: scsd8-2r-4 to 64 1838 Meszaros/sctap1-2b 3: sctap1-2b-4 to 64 1839 Meszaros/sctap1-2c 3: sctap1-2c-4 to 64 1840 Meszaros/sctap1-2r 11: sctap1-2r-4 to 480 Nov 9, 2006 * Recreated all MM and RB matrices from the MATLAB version Nov 8, 2006 * Added new fields to all Problems: Problem.date date matrix was created, or added to the collection if the creation date is unknown. Problem.author author of the matrix Problem.ed collector/editor of the matrix The fields are all strings. They are empty if not known, or if the date the matrix was added to the collection is very different than the probable creation date. * Added POLYFLOW and INPRO matrices from Parasol collection, collected by J. Koster. Note that most of the Parasol matrices are already in the collection, in the GHS_* sets. 1642 POLYFLOW/invextr1_new 1643 POLYFLOW/mixtank_new 1644 INPRO/msdoor Nov 7, 2006 * Added new matrices from the NEP collection (called the 'Bai' group here), id's 1612 to 1641. 1612 Bai/cryg10000 1613 Bai/cryg2500 1614 Bai/dw2048 1615 Bai/dw8192 1616 Bai/dwa512 1617 Bai/dwb512 1618 Bai/dwg961a 1619 Bai/dwg961b 1620 Bai/mhd1280a 1621 Bai/mhd1280b 1622 Bai/mhd3200a 1623 Bai/mhd3200b 1624 Bai/mhd4800a 1625 Bai/mhd4800b 1626 Bai/qh1484 1627 Bai/qh768 1628 Bai/rdb1250 1629 Bai/rdb1250l 1630 Bai/rdb200 1631 Bai/rdb200l 1632 Bai/rdb2048_noL see note below 1633 Bai/rdb3200l 1634 Bai/rdb450 1635 Bai/rdb450l 1636 Bai/rdb800l 1637 Bai/tols1090 1638 Bai/tols2000 1639 Bai/tols340 1640 Bai/tols4000 1641 Bai/tols90 * updated the following matrices: 314 Bai/mhda416 tiny (relative eps) errors in A 315 Bai/mhdb416 tiny (relative eps) errors in A 331 Bai/rdb2048 added note, below. Note added regarding Bai/rdb2048: This matrix (Bai/rdb2048) is the same as the original NEP RDB2048L matrix (as of Nov 2006). Either the name changed since 1996 and the L was added in the NEP collection, or the L was dropped when the matrix as added to the UF Sparse Matrix Collection in 1996. The name cannot be corrected here without disrupting the UF Sparse Matrix Collection. To make matters worse, there is now a matrix in the NEP collection with the name RDB2048, as of Nov 2006. The NEP RDB2048 matrix was added to this collection in Nov 2006, as Bai/rdb2038_noL matrix. Nov 6, 2006 * New matrices from the Cylshell set, Matrix Market (id's 1605 to 1611) Note that s3dkq4m2 and s3dkt3m2 are already in GHS_psdef. 1605 Cylshell/s1rmq4m1 1606 Cylshell/s2rmq4m1 1607 Cylshell/s3rmq4m1 1608 Cylshell/s1rmt3m1 1609 Cylshell/s2rmt3m1 1610 Cylshell/s3rmt3m1 1611 Cylshell/s3rmt3m3 * New matrices from the QCD set, Matrix Market (id's 1591 to 1604) The names were slightly modified, because (1) a "." in the name causes problems in the naming scheme in this collection and (2) they are long and unwieldly. id UF name original name ---- ------------------ ------------------- 1591 QCD/conf5_0-4x4-10 conf5.0-00l4x4-1000 1592 QCD/conf5_0-4x4-14 conf5.0-00l4x4-1400 1593 QCD/conf5_0-4x4-18 conf5.0-00l4x4-1800 1594 QCD/conf5_0-4x4-22 conf5.0-00l4x4-2200 1595 QCD/conf5_0-4x4-26 conf5.0-00l4x4-2600 1596 QCD/conf6_0-4x4-20 conf6.0-00l4x4-2000 1597 QCD/conf6_0-4x4-30 conf6.0-00l4x4-3000 1598 QCD/conf5_4-8x8-05 conf5.4-00l8x8-0500 1599 QCD/conf5_4-8x8-10 conf5.4-00l8x8-1000 1600 QCD/conf5_4-8x8-15 conf5.4-00l8x8-1500 1601 QCD/conf5_4-8x8-20 conf5.4-00l8x8-2000 1602 QCD/conf6_0-8x8-20 conf6.0-00l8x8-2000 1603 QCD/conf6_0-8x8-30 conf6.0-00l8x8-3000 1604 QCD/conf6_0-8x8-80 conf6.0-00l8x8-8000 * New matrices from Olaf Schenk added (id's 1537 to 1590) 1537 Schenk_IBMNA/c-18 1538 Schenk_IBMNA/c-19 1539 Schenk_IBMNA/c-20 1540 Schenk_IBMNA/c-21 1541 Schenk_IBMNA/c-22 1542 Schenk_IBMNA/c-23 1543 Schenk_IBMNA/c-24 1544 Schenk_IBMNA/c-25 1545 Schenk_IBMNA/c-26 1546 Schenk_IBMNA/c-27 1547 Schenk_IBMNA/c-28 1548 Schenk_IBMNA/c-29 1549 Schenk_IBMNA/c-30 1550 Schenk_IBMNA/c-31 1551 Schenk_IBMNA/c-32 1552 Schenk_IBMNA/c-33 1553 Schenk_IBMNA/c-34 1554 Schenk_IBMNA/c-35 1555 Schenk_IBMNA/c-36 1556 Schenk_IBMNA/c-37 1557 Schenk_IBMNA/c-38 1558 Schenk_IBMNA/c-39 1559 Schenk_IBMNA/c-40 1560 Schenk_IBMNA/c-41 1561 Schenk_IBMNA/c-42 1562 Schenk_IBMNA/c-43 1563 Schenk_IBMNA/c-44 1564 Schenk_IBMNA/c-45 1565 Schenk_IBMNA/c-46 1566 Schenk_IBMNA/c-47 1567 Schenk_IBMNA/c-48 1568 Schenk_IBMNA/c-49 1569 Schenk_IBMNA/c-50 1570 Schenk_IBMNA/c-51 1571 Schenk_IBMNA/c-52 1572 Schenk_IBMNA/c-53 1573 Schenk_IBMNA/c-54 1574 Schenk_IBMNA/c-56 1575 Schenk_IBMNA/c-57 1576 Schenk_IBMNA/c-60 1577 Schenk_IBMNA/c-61 1578 Schenk_IBMNA/c-65 1579 Schenk_IBMNA/c-big 1580 Schenk_AFE/af_0_k101 1581 Schenk_AFE/af_1_k101 1582 Schenk_AFE/af_2_k101 1583 Schenk_AFE/af_3_k101 1584 Schenk_AFE/af_4_k101 1585 Schenk_AFE/af_5_k101 1586 Schenk_AFE/af_shell10 1587 Schenk_IBMNA/c-64b 1588 Schenk_IBMNA/c-66b 1589 Schenk_IBMNA/c-67b 1590 Schenk_IBMNA/c-73b Note that c-64b, c-66b, c-67b, and c-73b are revised versions of c-64, c-66, c-67, and c-73, also in the Schenk_IBMNA group. The original matrices (c-64, etc) remain in the collection with their original names and id's. The four pairs ((c-64, c-64b), etc) placed in sequences. * GHS_indef/c-* problems updated. The following problems were obtained from Gould, Hu, and Scott, but originally came from IBM TJ Watson, via Olaf Schenk. (They ought to be in the Schenk_IBMNA directory, but that cannot now be changed since it would disrupt the use of these matrices by existing users of UFget). I downloaded the original S-CSR (compressed sparse row, S = symmetric) and RHS matrices from Olaf Schenk's site, which contains the original matrices: http://www.computational.unibas.ch/computer_science/scicomp/matrices and found that the matrices (nearly) match. c-62ghs is the same as the currently posted (Nov 2006) version of c-62 on Olaf Schenk's web site. However, the original matrices include right-hand-sides, which the GHS_indef matrices do not have. Thus, these were added to the matrices here. In addition, each of the matrices had tiny (1e-14) relative errors in the A matrix. The matrices were updated with the values posted at Schenk's web page. 1220 GHS_indef/c-55 1221 GHS_indef/c-58 1222 GHS_indef/c-59 1223 GHS_indef/c-62ghs 1224 GHS_indef/c-63 1225 GHS_indef/c-68 1226 GHS_indef/c-69 1227 GHS_indef/c-70 1228 GHS_indef/c-71 1229 GHS_indef/c-72 The pair of matrices (Schenk_IBMNA/c-62, GHS_indef/c-62ghs) placed in a sequence. Nov 2, 2006 * right-hand-side of Schenk_ISEI/barrier2-10 fixed. was wrong dimension. (original file was truncated when the Problem was first added to the Collection). Nov 1, 2006 * Additional duplicate matrices found. The column on the left lists the exact duplicates of the matrices on the right. For each duplicate matrix, the Problem.kind = 'duplicate' was added, and a Problem.note added, stating which matrix it's a duplicate of. The bcsstk* matrices are the same because the problems they reprsent have different mass matrices (bcsstm*). Torsion1 and obstclae come from different linear programs - same matrix A but different constraints on the variables (which are not present here). id name id name 29 HB/bcsstk07 see instead: 28 HB/bcsstk06 279 HB/young2c see instead: 278 HB/young1c 194 HB/lshp1009 see instead: 177 HB/jagmesh2 34 HB/bcsstk12 see instead: 33 HB/bcsstk11 1315 GHS_psdef/torsion1 see instead: 1314 GHS_psdef/obstclae 1241 GHS_indef/mario002 see instead: 1232 GHS_indef/darcy003 The following matrices in the left column are the binary patterns of the matrices in the right column. The matrices in the left column have been declared duplicates (Problem.kind = 'duplicate...' and Problem.notes added). 475 Gset/G15 is the pattern of 479 Gset/G19 477 Gset/G17 is the pattern of 482 Gset/G21 476 Gset/G16 is the pattern of 481 Gset/G20 474 Gset/G14 is the pattern of 478 Gset/G18 498 Gset/G36 is the pattern of 503 Gset/G40 497 Gset/G35 is the pattern of 501 Gset/G39 500 Gset/G38 is the pattern of 505 Gset/G42 499 Gset/G37 is the pattern of 504 Gset/G41 519 Gset/G55 is the pattern of 520 Gset/G56 522 Gset/G58 is the pattern of 523 Gset/G59 525 Gset/G60 is the pattern of 526 Gset/G61 528 Gset/G63 is the pattern of 529 Gset/G64 366 Boeing/pcrystk02 is the pattern of 351 Boeing/crystk02 367 Boeing/pcrystk03 is the pattern of 352 Boeing/crystk03 753 Mulvey/pfinan512 is the pattern of 752 Mulvey/finan512 733 Li/pli is the pattern of 732 Li/li Additional sequences of matrices already in the collection: len matrices --- -------- 4 DRIVCAV/cavity01 through 04 5 DRIVCAV/cavity05 through 09 6 DRIVCAV/cavity10 through 15 11 DRIVCAV/cavity16 through 26 Note that the nonzero patterns of matrices in the DRIVCAV sequences alternate. Every other matrix has a pattern that is the transpose of the pattern of the first matrix. Oct 31, 2006 * created sequence information for existing Problems (not merging into a single Problem struct). This format will be used for existing problems in the collection, with id's of 1378 or greater, not for new ones (unless new matrices are subsequent matrices in sequences started with new matrices). The first Problem in a sequence of MATLAB Problem struct has a Problem.kind = 'sequence'. The first line of Problem.notes for all problems in a sequence in the old format is a string of the form next: nextprob first: firstprob where nextprob is the name of the next Problem in the sequence, and firstprob is the name of the first Problem in the sequence. nextprob for the last Problem is '-'. Problems in a sequence, but not the first, have kind = 'subsequent'. This way, if you are interested in loading in just the primary matrices (not in a sequence, or not the first in a sequence), just avoid Problems with kind = 'subsequent'. One exception to this rule would be a sequence of basis matrices from an LP problem (bp_0, 200, etc). The matrices in those sequences are very different. Existing sequences handled in this manner: len Problems --- -------- 70 Sandia/init_adder1, and Sandia/adder_dcop_01 through _69 2 Sandia/adder_trans_01 through _02 51 Sandia/fpga_dcop_01 through _51 2 Sandia/fpga_trans_01 through _02 3 Sandia/mult_dcop_01 through _03 57 Sandia/oscil_dcop_01 through _57 26 Nemeth/nemeth01 through 26 9 Schenk_AFE/af_shell1 through 9 3 Simon/venkat01, Simon/venkat25, Simon/venkat50 2 Simon/raefsky1, Simon/raefsky2 9 HB/bp_0, _200, _400, _600, _800, _1000, _1200, _1400, _1600 2 HB/gemat11, 12 3 HB/shl_0, _200, _400 4 HB/str_0, _200, _400, _600 4 HB/fs_183_1, _3, _4, _6 4 HB/fs_541_1, _2, _3, _4 3 HB/fs_680_1, _2, _3 3 HB/fs_760_1, _2, _3 6 Bai/cdde1 through 6 2 Wang/swang1, 2 2 Wang/wang1, 2 4 Schenk_ISEI/barrier2-1, 2, 3, 4 4 Schenk_ISEI/barrier2-9, 10, 11, 12 6 Schenk_ISEI/para-5, 6, 7, 8, 9, 10 (para-4 has different size) 3 IBM_EDA/dc1, 2, 3 2 IBM_EDA/trans4, 5 New sequences will be added using the aux.cell array format, and a single Matrix Market directory. The above sequences all appear in their own Problem struct, and their own Matrix Market directory. The name of the cell array is problem dependent, but all cell arrays in the aux struct will always contain the sequence itself. Problem.kind = 'sequence' will be used for each Problem containing a sequence, but there will be no Problem.kind = 'subsequent'. Sequences in new format as of Oct 31, 2006: len Problems --- -------- 44 Zitney/extr1b aux.A{1:44}, aux.Zeros{1:44} 76 Zitney/hydr1c aux.A{1:76}, aux.Zeros{1:76} 66 Reuters911 aux.Days{1:66} The old format is used exclusively for Problem id's of 1377 and less (matrices added May 2006 and earlier). The new format will be be used for matrices added after that date, with Problem id's of 1378 or greater. * There are some duplicate matrices, or near duplicates, in the collection. The following matrices have identical name, but are not in the same groups. These matrices are all related. For a matrix declared 'duplicate', a Problem.kind = 'duplicate' has been added. The following have been declared duplicate: Nasa/barth see Pothen/barth instead Nasa/barth4 see Pothen/barth4 instead Nasa/barth5 see Pothen/barth5 instead Boeing/nasa1824 see Nasa/nasa1824 Boeing/nasa2910 see Nasa/nasa2910 Boeing/nasa4704 see Nasa/nasaa4704 Nasa/shuttle_eddy see Pothen/shuttle_eddy instead Nasa/skirt see Pothen/skirt instead GHS_psdef/copter2 see GHS_indef/copter2 instead Nasa/pwt see Pothen/pwt instead GHS_psdef/pwt see Pothen/pwt instead Details: A1=Nasa/barth, A2=Pothen/barth. both are binary, A2 = spones(A1+A1'). A2 has xyz coordinates. A1 declared duplicate. A1=Nasa/barth4, A2=Pothen/barth4. both are binary, A2 = spones(A1+A1'). A2 has xyz coordinates. A1 declared duplicate. A1=Nasa/barth5, A2=Pothen/barth5. both are binary, A2 = spones(A1+A1'). A2 has xyz coordinates. A1 declared duplicate. A1=Nasa/nasa1824, A2=Boeing/nasa1824. both have values, same pattern, different values: 386 out of 39208 in small area of the matrix, tril (A(196:321, 196:216)), and the same in triu (A (...)). The magnitudes of the A2 entries that differ from A1 are very small (~1e-6 or ~1e-9), unlike most entries in the rest of the matrix. Also, the A2 entries are single precision values (only 9 digits when printed in base 10), unlike the other entries in A1 and A2. A1 has entries of more similar magnitude. I suspect A2 might be a corrupted version of A1. A2 declared duplicate. Do a "diff" of the Matrix Market files and you will see why I conclude this. A1=Nasa/nasa2910, A2=Boeing/nasa2910. A1 has values, A2 = spones (A1). A2 declared duplicate. A1=Nasa/nasa4704, A2=Boeing/nasa4704. A1 has values, A2 = spones (A1). A2 declared duplicate. A1=Nasa/shuttle_eddy, A2=Pothen/shuttle_eddy. both binary, A1 = A2, A2 has xyz coordinates A1=Nasa/skirt, A2=Pothen/skirt. both binary, A1 = A2, A2 has xyz coords A1 declated duplicate. A1=GHS_indef/copter2, A2=GHS_psdef/copter2. A1 has values, A2 = spones (A1). A2 declared duplicate. A1=Nasa/pwt, A2=GHS_psdef/pwt, A3=Pothen/pwt, all are binary, A1 = A2 = A3, A3 has xyz coordinates. A1 and A2 declared duplicate. In some cases, the duplicates are intentional. One a matrix name and id is entered in UF_Listing.txt, it is never modified. The Nasa matrices were added first, but then Alex Pothen obtained versions with xyz coordinates. Both versions were kept in the collection. In other cases, the duplicates are a mistake. Rather than removing them, I've chosen to keep to my rule of leaving the UF_Listing.txt (mapping of Problem names to Problem id's) unchanged. I only append to that list. The following matrices (id 1202 to 1211) are also intentionally duplicated, because I have restructured how these problems are represented. Many more Oberwolfach model reduction problems have been added, and it made more sense to restructure the collection so that each Oberwolfach problem appears as one MATLAB Problem struct and as one Matrix Market directory. id name ---- ----- 1202 Oberwolfach/gyro_k 1203 Oberwolfach/gyro_m 1204 Oberwolfach/t2dah_a 1205 Oberwolfach/t2dah_e 1206 Oberwolfach/t2dal_a 1207 Oberwolfach/t2dal_e 1208 Oberwolfach/t3dh_a 1209 Oberwolfach/t3dh_e 1210 Oberwolfach/t3dl_a 1211 Oberwolfach/t3dl_e in the new format, all matrices for each problem (A, K, M, E, B, C, etc) all appear in a single Problem struct: id name ---- ----- 1435 Oberwolfach/gyro 1447 Oberwolfach/t2dah 1449 Oberwolfach/t2dal 1450 Oberwolfach/t3dh 1451 Oberwolfach/t3dl For example, the gyro_k matrix thus appears as both 1202 Oberwolfach/gyro_k 1435 Oberwolfach/gyro These two copies of the gyro_k matrix are identical. gyro_k appears just once in the Harwell/Boeing format of the collection as an individual file (gyro_k.rsa). Oct 30, 2006 * Added Zitney sequences (problems 1535 and 1536) Zitney/extr1b 44 matrices Zitney/hydr1c 76 matrices These two MATLAB problems are sequences of 44 and 76 matrices arising in a nonlinear solver. Each sequence appears in a single MATLAB Problem struct, with kind = 'sequence', and in a single Matrix Market *.tar.gz file. The sequences are expanded as individual files in the Harwell/Boeing form of the collection. A sequence of graphs use kind = '... graph sequence' instead of just kind = 'sequence'. In both cases, the word 'sequence' can be found in Problem.kind. All matrices in a sequence share a single id. They appear in a Problem.aux.whatever{...} cell array in the MATLAB version. This method will be used for sequences added to the collection from now on. Existing sequences already in the collection (Sandia, Nemeth) will be handled differently. * Added Problem.kind to the following MATLAB problems: LPnetlib/*, Qaplib/* : 'linear programming problem' Gset/*: 'undirected random graph' or 'undirected weighted random graph' Simon/appu: 'random matrix' Note that appu is very nearly a symmetric matrix (norm (A-A') is eps). * Added Pajek networks, problems 1456 to 1534 Pajek data sets, from http://vlado.fmf.uni-lj.si/pub/networks/data/, Vladimir Batagelj and Andrej Mrvar (2006): Pajek datasets. Creative Commons Attribution-NonCommercial-ShareAlike 2.5 License. Pajek/California Pajek/Cities Pajek/CSphd Pajek/dictionary28 Pajek/divorce Pajek/EAT_RS Pajek/EAT_SR Pajek/EPA Pajek/Erdos02 Pajek/Erdos971 Pajek/Erdos972 Pajek/Erdos981 Pajek/Erdos982 Pajek/Erdos991 Pajek/Erdos992 Pajek/EVA Pajek/FA Pajek/foldoc Pajek/football Pajek/GD00_a Pajek/GD00_c Pajek/GD01_a Pajek/GD01_A Pajek/GD01_b Pajek/GD01_c Pajek/GD02_a Pajek/GD02_b Pajek/GD06_Java Pajek/GD06_theory Pajek/GD95_a Pajek/GD95_b Pajek/GD95_c Pajek/GD96_a Pajek/GD96_b Pajek/GD96_c Pajek/GD96_d Pajek/GD97_a Pajek/GD97_b Pajek/GD97_c Pajek/GD98_a Pajek/GD98_b Pajek/GD98_c Pajek/GD99_b Pajek/GD99_c Pajek/geom Pajek/GlossGT Pajek/HEP-th Pajek/HEP-th-new Pajek/IMDB Pajek/internet Pajek/Journals Pajek/Kohonen Pajek/Lederberg Barabasi/NotreDame_actors Barabasi/NotreDame_www Barabasi/NotreDame_yeast Pajek/ODLIS Pajek/patents_main Pajek/patents Pajek/Ragusa16 Pajek/Ragusa18 Pajek/Reuters911 kind = 'undirected weighted graph sequence' Pajek/Roget Pajek/Sandi_authors Pajek/Sandi_sandi Pajek/SciMet Pajek/SmaGri Pajek/SmallW Pajek/Stranke94 Pajek/Tina_AskCal Pajek/Tina_AskCog Pajek/Tina_DisCal Pajek/Tina_DisCog Pajek/USAir97 Pajek/USpowerGrid Pajek/Wordnet3 Pajek/WorldCities Pajek/yeast Pajek/Zewail * Restructured MATLAB Problems The fields in the MATLAB versions of the problems have been simplified. All MATLAB problems have the following fields (this has not changed): name the group directory and problem name (i.e. HB/arc130) title short descriptive title A an m-by-n sparse matrix id an integer in the range of 1 to the # of problems in the collection. Identical to the line number of http://www.cise.ufl.edu/research/sparse/mat/UF_Listing.txt New matrices are always added at the end of this list. A MATLAB Problem may also have the following optional fields. The first three have not changed, the latter three are new. Zeros pattern of explicit zero entries in A, as a binary matrix. These entries are provided by the matrix submittor with zero numerical value. b right-hand-side of size n-by-(whatever) x supposed solution to A*x=b kind a string (i.e. 'directed graph') notes a character array with notes about the problem aux a struct, containing auxiliary information. The contents of this struct is problem dependent. Typical information in aux includes: colname: colname(j,:) is the name of column j rowname: rowname(i,:) is the name of row i nodename: nodename(i,:) is the name of row and column i coord: an n-by-2 or n-by-3 array of xy or xyz coordinates of each row/column/node of the matrix c: constraint vector for a linear programming (LP) problem lo: lower bounds for an LP hi: upper boundw for an LP z0: initial value of the objective for an LP E, M, B, C: additional matrices for model reduction problem cnames: name of rows of C for model reduction problem There are many other possible fields to "aux". For example, Pajek/Reuters911 has aux.Days, which is a cell array of 66 sparse matrices, one for each day (Problem.A is the sum of all 66 aux.Days{i} matrices). Most aux data is information about each row/column of the matrix (such as category, date, etc.). Changes made accordingly: LPnetlib/*: moved c, lo, hi, z0 to aux.c, aux.lo, aux.hi, aux.z0 Pothen/*: moved coord to aux.coord Engwirda/airfoil_2d: moved coord and t to aux.coord and aux.t Bindel/ted_AB, ted_AB_unscaled: moved w0 to aux.w0 MathWorks/Harvard500: changed cell colname to char array aux.nodename added kind = 'directed graph' Kamvar/Stanford_Berkeley, Kamvar/Stanford Problem.colname removed. Problem.aux.roots and Problem.aux.rootname added in its place. These are the 466 root nodes for the web crawl that generated the matrix. The matrix is unchanged. Added kind = 'directed graph'. * Matrix Market files reconstructed. Similar format (gzip'ed tar files), except that the files now all have a ".mtx" extension. Oct 25, 2006 * line counts for some complex matrices were incorrect in Harwell/Boeing format. Harwell/Boeing matrices regenerated for the following: Bindel/ted_AB.cua Bindel/ted_AB_unscaled.cua Cote/mplate.csa FEMLAB/waveguide3D.cua Kim/kim1.cua Kim/kim2.cua Oberwolfach/windscreen_k.csa Okunbor/aft02.cua * DNVS/*.rse.gz matrices were in DOS file format (trailing control-M's); converted to UNIX format. DNVS/m_t1.rse DNVS/ship_001.rse DNVS/ship_003.rse DNVS/shipsec1.rse DNVS/shipsec5.rse DNVS/shipsec8.rse DNVS/x104.rse * key field in some matrices in Harwell/Boeing format was lower case (rsa instead of RSA). These should all be upper case. The following files were corrected. DNVS/* files were in DOS format as well; converted to UNIX format. DNVS/m_t1.rsa DVNS/ship_003.rsa DNVS/shipsec1.rsa DNVS/shipsec5.rsa DNVS/shipsec8.rsa DNVS/thread.rsa DNVS/x104.rsa GHS_indef/bloweybl.rsa GHS_indef/bloweybq.rsa GHS_indef/bmw3_2.rsa GHS_indef/copter2.rsa GHS_indef/cvxqp3.rsa GHS_indef/darcy003.rsa GHS_indef/dawson5.rsa GHS_indef/d_pretok.rsa GHS_indef/exdata_1.rsa GHS_indef/helm2d03.rsa GHS_indef/helm2d01.rsa GHS_indef/k1_san.rsa GHS_indef/mario001.rsa GHS_indef/mario002.rsa GHS_indef/olesnik0.rsa GHS_indef/sit100.rsa GHS_indef/turon_m.rsa GHS_psdef/audikw_1.rsa GHS_psdef/bmw7st_1.rsa GHS_psdef/bmwcra_1.rsa GHS_psdef/crankseg_1.rsa GHS_psdef/crankseg_2.rsa GHS_psdef/hood.rsa GHS_psdef/inline_1.rsa GHS_psdef/ldoor.rsa GHS_psdef/oilpan.rsa GHS_psdef/s3dkq4m2.rsa GHS_psdef/s3dkt3m2.rsa GHS_psdef/vanbody.rsa * Reconstructed HB/young*c matrices (both HB format and MATLAB). young3c now appears as an RUA matrix. The YOUNG*C matrices originally appeared in the Harwell/Boeing collection as type CSA (complex symmetric). However, both upper and lower triangular parts are present in the original files (an invalid specification; only the lower part can be present in the file). If the entries in the upper triangular part are considered as part of the matrix, the matrices become unsymmetric. The matrices have been corrected in the UF Sparse Matrix Collection by changing their type to CUA so that the entries in the original files are not ignored. In addition, the YOUNG3C matrix has a zero imaginary part, and thus appears here as a real matrix. Oct 20, 2006 * Boeing/bcsstk39.rsa reconstructed. The original file had one extra entry past the last nonzero. * added new matrices (id's 1529 to 1534) GHS_psdef/apache1 GHS_psdef/apache2 GHS_indef/bloweybl GHS_indef/bloweybq GHS_indef/cvxqp3 GHS_indef/qpband The original source was in the Harwell/Boeing format, but the apache1, apache2, and qpband matrices had unsorted columns and/or duplicate row indices. These are correct in my Harwell/Boeing versions (duplicates are summed and columns are sorted). * MAJOR CHANGE: Harwell/Boeing format of LiuWnzhou/powersim.rua matrix was incorrect. In that file, the numerical values of the matrix were all equal to one. The MATLAB version was correct; HB format regenerated. * MAJOR CHANGE: Harwell/Boeing format of Engwirda/airfile_2d.rsa matrix was incorrect. The matrix is unsymmetric, not symmetric. The MATLAB version was correct; HB format regenerated (as airfoil_2d.rua, replacing airfoil_2d.rsa). Oct 19, 2006 * Oberwolfach matrices regenerated from original sources (gyro, t2da*, t3da*), many new matrices added. Oct 18, 2006 * Note added to Bydder/* matrices, showing how to construct the MRI images in MATLAB. * MAJOR CHANGE: correction to Harwell/Boeing format of IBM_Austin/coupled. The matrix is not symmetric. coupled.rsa deleted, the correct coupled.rua created in its place. * tiny changes (relative eps) to Problem.x in MATLAB files, to match Harwell/Boeing: x error 1.8e-16 DRIVCAV/cavity25 x error 1.4e-16 DRIVCAV/cavity24 x error 1.2e-16 Simon/bbmat * tiny changes (relative eps) to Problem.b in MATLAB files, to match Harwell/Boeing: b error 1.2e-16 Simon/venkat25 b error 1.9e-16 Simon/venkat50 * tiny (relative epsilon) changes to Problem.A matrices in MATLAB files, to match Harwell/Boeing files: A error 1.3e-16 FIDAP/ex2 A error 1.4e-16 FIDAP/ex10 A error 5.4e-16 GHS_indef/ncvxqp9 A error 5.7e-16 GHS_indef/ncvxqp1 A error 5.4e-16 GHS_indef/aug3dcqp A error 3.3e-16 GHS_indef/stokes64 A error 5.2e-16 GHS_indef/stokes64s A error 3.4e-16 GHS_indef/mario001 A error 5.6e-16 GHS_indef/helm3d01 A error 3.3e-16 GHS_indef/stokes128 A error 3.4e-16 GHS_indef/darcy003 A error 3.4e-16 GHS_indef/mario002 A error 5.2e-16 GHS_indef/exdata_1 A error 5.1e-16 GHS_indef/helm2d03 A error 5.6e-16 GHS_psdef/wathen100 A error 5.7e-16 GHS_psdef/wathen120 A error 5.7e-16 GHS_psdef/vanbody A error 5.6e-16 GHS_psdef/s3dkt3m2 A error 5.7e-16 GHS_psdef/s3dkq4m2 A error 5.7e-16 GHS_psdef/bmw7st_1 A error 5.7e-16 GHS_psdef/bmwcra_1 A error 5.7e-16 GHS_psdef/inline_1 A error 1.1e-16 DNVS/shipsec8 A error 1.1e-16 DNVS/shipsec1 A error 1.1e-16 DNVS/ship_003 A error 2.2e-16 Simon/raefsky3 A error 2.1e-16 Simon/bbmat A error 1.6e-16 Nasa/nasa2910 A error 1.8e-16 Boeing/pwtk * PSA and RSA matrices added to Harwell/Boeing format, constructed for all PSE and RSE matrices that did not already have them. HB/cegb2802.psa HB/cegb2919.psa HB/cegb3024.psa HB/cegb3306.psa HB/lap_25.psa HB/lock1074.psa HB/lock2232.psa HB/lock3491.psa HB/lock_700.psa HB/man_5976.psa GHS_psdef/opt1.psa GHS_psdef/ramage02.psa GHS_psdef/srb1.psa DNVS/crplat2.psa DNVS/fcondp2.psa DNVS/fullb.psa DNVS/halfb.psa DNVS/trdheim.psa DNVS/troll.psa DNVS/tsyl201.psa DNVS/ship_001.rsa DNVS/m_t1.rsa already in DNVS DNVS/ship_003.rsa already in DNVS DNVS/shipsec1.rsa already in DNVS DNVS/shipsec5.rsa already in DNVS DNVS/shipsec8.rsa already in DNVS DNVS/thread.rsa already in DNVS DNVS/x104.rsa already in DNVS Comparing my created RSA matrices with pre-existing RSA versions matrices in DNVS collection (from Gould, Hu, and Scott): name error (relative componentwise) m_t1 0 ship003 1e-16 shipsec1 1e-16 shipsec5 0 shipsec8 1e-16 thread 0 x104 0 * Renamed Harwell/Boeing files, to match MATLAB version of collection: from: to: Mulvey/finan512.psa Mulvey/pfinan512.psa Boeing/crystk02.psa Boeing/pcrystk02.psa Boeing/crystk03.psa Boeing/pcrystk03.psa Boeing/ct20stif.psa Boeing/pct20stif.psa Li/li.psa Li/pli.psa Note that the p*.psa matrices are nonzero patterns of related *.rsa matrices. This change did not modify UF_Listing.txt; only the HB file name was modified to match the UF_Listing.txt file. Oct 17, 2006 * VanVelzen HB files did not have the explicit zeros. Added from MALTAB files. * Rajat MATLAB and HB files recreated. MATLAB files were fine, just with nzmax > nnz. The new files lead to lower memory usage in MATLAB when read in. The old HB files were generated from scratch. * MAJOR CHANGE: renamed Harwell/Boeing matrices young*c.csa to young*c.cua. The files contain entries in the upper triangular part, which breaks the "csa" standard (entries in symmetric H/B file must only appear in the lower triangular part). Note that the imaginary part of young3c.cua is all zero; this could have been stored in the collection as an 'rua' matrix. * Alemdar/Alemdar.rsa and Andrews/Andrews.rsa fixed. The old versions had both upper & lower parts present. * Reconstructed IBM_EDA/dc* and trans* RUA files (explicit zero pattern missing from RUA files) Oct 16, 2006 * MAJOR CHANGE: Hamrle*.rua files corrected. The right-hand-sides were wrong. * Sandia/oscil*, Sandia/fpga*, Sandia/mult* Harwell/Boeing files updated. right-hand-sides added. * MathWorks/pivtol.rua, Pd.rua updated. RHS added from MATLAB version. * Added Problem.Zeros to Lucifora/cell1 and cell2 MATLAB files October, 2006 * New directories: Andrianov MKS Raju Bates Schmid AMD Bindel Koutsovasilis Oberwolfach * New matrices (id's 1378 to 1455), from The MathWorks, Alexander Andrianov, Dan Smith (MKS), Rajat, Mandhapati Raju, Doug Bates, Dani Schmid, David Bindel, Panagiotis Koutsovasilis, AMD Inc. (Ufuk Okuyucu), Sandia National Labs, IBM (EDA), and the Oberwolfach collection. MathWorks/Sieber Andrianov/ex3sta1 Andrianov/fxm3_6 Andrianov/fxm4_6 Andrianov/ins2 Andrianov/lp1 Andrianov/lpl1 Andrianov/mip1 Andrianov/net100 Andrianov/net125 Andrianov/net150 Andrianov/net25 Andrianov/net4-1 Andrianov/net50 Andrianov/net75 Andrianov/pattern1 Andrianov/pf2177 MKS/fp Rajat/rajat29 Rajat/rajat30 Rajat/rajat31 Raju/laminar_duct3D Bates/Chem97Zt Bates/Chem97ZtZ Schmid/thermal1 Schmid/thermal2 MathWorks/Kaufhold Bindel/ted_A Bindel/ted_B Bindel/ted_AB Bindel/ted_A_unscaled Bindel/ted_B_unscaled Bindel/ted_AB_unscaled Koutsovasilis/F1 AMD/G2_circuit IBM_EDA/ckt11752_dc_1 IBM_EDA/ckt11752_tr_0 Sandia/ASIC_100k Sandia/ASIC_100ks Sandia/ASIC_320k Sandia/ASIC_320ks Sandia/ASIC_680k Sandia/ASIC_680ks AMD/G3_circuit GHS_psdef/apache1 GHS_psdef/apache2 GHS_indef/bloweybl GHS_indef/bloweybq GHS_indef/cvxqp3 GHS_indef/qpband Oberwolfach/chipcool0 Oberwolfach/chipcool1 Oberwolfach/filter2D Oberwolfach/filter3D Oberwolfach/flowmeter0 Oberwolfach/flowmeter5 Oberwolfach/gas_sensor Oberwolfach/gyro Oberwolfach/inlet Oberwolfach/LF10000 Oberwolfach/LF10 Oberwolfach/LFAT5000 Oberwolfach/LFAT5 Oberwolfach/piston Oberwolfach/rail_1357 Oberwolfach/rail_20209 Oberwolfach/rail_5177 Oberwolfach/rail_79841 Oberwolfach/spiral Oberwolfach/t2dah Oberwolfach/t2dal_bci Oberwolfach/t2dal Oberwolfach/t3dh Oberwolfach/t3dl Oberwolfach/windscreen Oberwolfach/bone010 Oberwolfach/boneS01 Oberwolfach/boneS10 May 2006: * MatrixMarket version of the collection created (matrices 1 to 1377) March 2006: * New directories Bydder Engwirda IBM_EDA Morandini Toledo VanVelzen Kemelmacher Lourakis PARSEC * New matrices (id's 1316 to 1377), from Rajat, Mark Bydder, Daniel Engwirda, IBM, Marco Morandini, The MathWorks, Sivan Toledo, Nils van Velzen, M. Lourakis, and the PARSEC problems (Yunkai Zhou and colleagues) Rajat/rajat15 Bydder/mri1 Bydder/mri2 Engwirda/airfoil_2d IBM_EDA/dc1 IBM_EDA/dc2 IBM_EDA/dc3 IBM_EDA/trans4 IBM_EDA/trans5 Morandini/robot Morandini/rotor1 Morandini/rotor2 MathWorks/tomography Kemelmacher/Kemelmacher MathWorks/Kuu MathWorks/Muu Toledo/deltaX VanVelzen/std1_Jac2_db VanVelzen/std1_Jac2 VanVelzen/std1_Jac3_db VanVelzen/std1_Jac3 VanVelzen/Zd_Jac2_db VanVelzen/Zd_Jac2 VanVelzen/Zd_Jac3_db VanVelzen/Zd_Jac3 VanVelzen/Zd_Jac6_db VanVelzen/Zd_Jac6 Rajat/rajat16 Rajat/rajat17 Rajat/rajat18 Rajat/rajat19 Lourakis/bundle1 PARSEC/benzene PARSEC/CO PARSEC/Ga10As10H30 PARSEC/Ga19As19H42 PARSEC/Ga3As3H12 PARSEC/Ga41As41H72 PARSEC/GaAsH6 PARSEC/Ge87H76 PARSEC/Ge99H100 PARSEC/H2O PARSEC/Na5 PARSEC/Si10H16 PARSEC/Si2 PARSEC/Si34H36 PARSEC/Si41Ge41H72 PARSEC/Si5H12 PARSEC/Si87H76 PARSEC/SiH4 PARSEC/SiNa PARSEC/SiO2 PARSEC/SiO Rajat/rajat20 Rajat/rajat21 Rajat/rajat22 Rajat/rajat23 Rajat/rajat24 Rajat/rajat25 Rajat/rajat26 Rajat/rajat27 Rajat/rajat28 August 2004: * updated Mathworks: added Pd* and Harvard500 matrices * updated HB: added *.pse matrices to MATLAB collection * added new directories: IBM_Austin ACUSIM GHS_psdef GHS_indef Rajat Hamrle Oberwolfach LiuWenzhuo Lin Cannizzo * added 146 matrices, id's (1170 to 1316), from The MathWorks, IBM (Austin), ACUSIM, Rajat, Jaroslav Hamrle, the Oberwolfach collection, Wenzhou Liu, Martin Lin, Fabio Connizz, and a large suite (GHS) collected by Nick Gould, Yifan Hu, and Jennifer Scott (including the DNVS matrices). The HB matrices are assembled versions of unassembled finite-element matrices already in the HB collection (these were added to the MATLAB collection, but the Harwell/Boeing format was left in PSE and RSE formats). Note that some of original GHS matrices contain unsorted columns MathWorks/Pd MathWorks/Pd_rhs MathWorks/Harvard500 HB/cegb2802 HB/cegb2919 HB/cegb3024 HB/cegb3306 HB/lap_25 HB/lock1074 HB/lock2232 HB/lock3491 HB/lock_700 HB/man_5976 IBM_Austin/coupled ACUSIM/Pres_Poisson Rajat/rajat01 Rajat/rajat02 Rajat/rajat03 Rajat/rajat04 Rajat/rajat05 Rajat/rajat06 Rajat/rajat07 Rajat/rajat08 Rajat/rajat09 Rajat/rajat10 Rajat/rajat11 Rajat/rajat12 Rajat/rajat13 Rajat/rajat14 Hamrle/Hamrle1 Hamrle/Hamrle2 Hamrle/Hamrle3 Oberwolfach/gyro_k Oberwolfach/gyro_m Oberwolfach/t2dah_a Oberwolfach/t2dah_e Oberwolfach/t2dal_a Oberwolfach/t2dal_e Oberwolfach/t3dh_a Oberwolfach/t3dh_e Oberwolfach/t3dl_a Oberwolfach/t3dl_e LiuWenzhuo/powersim Lin/Lin Cannizzo/sts4098 GHS_indef/aug2d GHS_indef/aug2dc GHS_indef/aug3d GHS_indef/aug3dcqp GHS_indef/bmw3_2 GHS_indef/c-55 GHS_indef/c-58 GHS_indef/c-59 GHS_indef/c-62ghs GHS_indef/c-63 GHS_indef/c-68 GHS_indef/c-69 GHS_indef/c-70 GHS_indef/c-71 GHS_indef/c-72 GHS_indef/copter2 GHS_indef/d_pretok GHS_indef/darcy003 GHS_indef/dawson5 GHS_indef/dtoc GHS_indef/exdata_1 GHS_indef/helm2d03 GHS_indef/helm3d01 GHS_indef/k1_san (original file not terminated with newline) GHS_indef/laser GHS_indef/mario001 GHS_indef/mario002 GHS_indef/ncvxqp1 GHS_indef/ncvxqp9 GHS_indef/olesnik0 (original file not terminated with newline) GHS_indef/sit100 (original file not terminated with newline) GHS_indef/stokes128 GHS_indef/stokes64 GHS_indef/stokes64s GHS_indef/tuma1 GHS_indef/tuma2 GHS_indef/turon_m GHS_psdef/audikw_1 GHS_psdef/bmw7st_1 GHS_psdef/bmwcra_1 GHS_psdef/copter1 GHS_psdef/copter2 GHS_psdef/crankseg_1 GHS_psdef/crankseg_2 DNVS/crplat2 DNVS/fcondp2 GHS_psdef/finance256 GHS_psdef/ford1 GHS_psdef/ford2 DNVS/fullb DNVS/halfb GHS_psdef/hood GHS_psdef/inline_1 GHS_psdef/ldoor DNVS/m_t1 GHS_psdef/oilpan GHS_psdef/opt1 GHS_psdef/pds10 GHS_psdef/pwt GHS_psdef/ramage02 GHS_psdef/s3dkq4m2 GHS_psdef/s3dkt3m2 DNVS/ship_001 DNVS/ship_003 DNVS/shipsec1 DNVS/shipsec5 DNVS/shipsec8 GHS_psdef/srb1 DNVS/thread DNVS/trdheim DNVS/troll DNVS/tsyl201 GHS_psdef/vanbody GHS_psdef/wathen100 GHS_psdef/wathen120 DNVS/x104 GHS_indef/a0nsdsil (unsorted cols and/or duplicate entries) GHS_indef/a2nnsnsl " GHS_indef/a5esindl " GHS_indef/blockqp1 " GHS_indef/bloweya " GHS_indef/boyd1 " GHS_indef/boyd2 " GHS_indef/brainpc2 " GHS_indef/bratu3d " GHS_indef/cont-201 " GHS_indef/cont-300 " GHS_indef/dixmaanl " GHS_indef/linverse " GHS_indef/ncvxbqp1 " GHS_indef/ncvxqp3 " GHS_indef/ncvxqp5 " GHS_indef/ncvxqp7 " GHS_indef/sparsine " GHS_indef/spmsrtls " GHS_psdef/cvxbqp1 " GHS_psdef/gridgena " GHS_psdef/jnlbrng1 " GHS_psdef/minsurfo " GHS_psdef/obstclae " GHS_psdef/torsion1 " October, 2003 * added new matrices (id's 923 to 1169), from Bulent Alemdar, Stuart Andrews, Comsol Inc. (FEMLAB), Salvatore Lucifora, the ND set, Olaf Schenk (collected from various sources), Sep Kamvar, Bob Sumner, Roy Tromble, Jagesh Sanghavi, and Sandia National Labs Alemdar/Alemdar Andrews/Andrews FEMLAB/ns3Da FEMLAB/poisson2D FEMLAB/poisson3Da FEMLAB/poisson3Db FEMLAB/problem1 FEMLAB/sme3Da FEMLAB/sme3Db FEMLAB/sme3Dc FEMLAB/waveguide3D Lucifora/cell1 Lucifora/cell2 ND/nd3k ND/nd6k ND/nd12k ND/nd24k Schenk_AFE/af_shell1 Schenk_AFE/af_shell2 Schenk_AFE/af_shell3 Schenk_AFE/af_shell4 Schenk_AFE/af_shell5 Schenk_AFE/af_shell6 Schenk_AFE/af_shell7 Schenk_AFE/af_shell8 Schenk_AFE/af_shell9 Schenk_IBMNA/c-62 Schenk_IBMNA/c-64 Schenk_IBMNA/c-66 Schenk_IBMNA/c-67 Schenk_IBMNA/c-73 Schenk_IBMSDS/2D_27628_bjtcai Schenk_IBMSDS/2D_54019_highK Schenk_IBMSDS/3D_28984_Tetra Schenk_IBMSDS/3D_51448_3D Schenk_IBMSDS/ibm_matrix_2 Schenk_IBMSDS/matrix_9 Schenk_IBMSDS/matrix-new_3 Schenk_ISEI/barrier2-10 Schenk_ISEI/barrier2-11 Schenk_ISEI/barrier2-12 Schenk_ISEI/barrier2-1 Schenk_ISEI/barrier2-2 Schenk_ISEI/barrier2-3 Schenk_ISEI/barrier2-4 Schenk_ISEI/barrier2-9 Schenk_ISEI/igbt3 Schenk_ISEI/nmos3 Schenk_ISEI/ohne2 Schenk_ISEI/para-10 Schenk_ISEI/para-4 Schenk_ISEI/para-5 Schenk_ISEI/para-6 Schenk_ISEI/para-7 Schenk_ISEI/para-8 Schenk_ISEI/para-9 Kamvar/Stanford Kamvar/Stanford_Berkeley Sumner/graphics Tromble/language Sanghavi/ecl32 Sandia/adder_dcop_01 Sandia/adder_dcop_02 Sandia/adder_dcop_03 Sandia/adder_dcop_04 Sandia/adder_dcop_05 Sandia/adder_dcop_06 Sandia/adder_dcop_07 Sandia/adder_dcop_08 Sandia/adder_dcop_09 Sandia/adder_dcop_10 Sandia/adder_dcop_11 Sandia/adder_dcop_12 Sandia/adder_dcop_13 Sandia/adder_dcop_14 Sandia/adder_dcop_15 Sandia/adder_dcop_16 Sandia/adder_dcop_17 Sandia/adder_dcop_18 Sandia/adder_dcop_19 Sandia/adder_dcop_20 Sandia/adder_dcop_21 Sandia/adder_dcop_22 Sandia/adder_dcop_23 Sandia/adder_dcop_24 Sandia/adder_dcop_25 Sandia/adder_dcop_26 Sandia/adder_dcop_27 Sandia/adder_dcop_28 Sandia/adder_dcop_29 Sandia/adder_dcop_30 Sandia/adder_dcop_31 Sandia/adder_dcop_32 Sandia/adder_dcop_33 Sandia/adder_dcop_34 Sandia/adder_dcop_35 Sandia/adder_dcop_36 Sandia/adder_dcop_37 Sandia/adder_dcop_38 Sandia/adder_dcop_39 Sandia/adder_dcop_40 Sandia/adder_dcop_41 Sandia/adder_dcop_42 Sandia/adder_dcop_43 Sandia/adder_dcop_44 Sandia/adder_dcop_45 Sandia/adder_dcop_46 Sandia/adder_dcop_47 Sandia/adder_dcop_48 Sandia/adder_dcop_49 Sandia/adder_dcop_50 Sandia/adder_dcop_51 Sandia/adder_dcop_52 Sandia/adder_dcop_53 Sandia/adder_dcop_54 Sandia/adder_dcop_55 Sandia/adder_dcop_56 Sandia/adder_dcop_57 Sandia/adder_dcop_58 Sandia/adder_dcop_59 Sandia/adder_dcop_60 Sandia/adder_dcop_61 Sandia/adder_dcop_62 Sandia/adder_dcop_63 Sandia/adder_dcop_64 Sandia/adder_dcop_65 Sandia/adder_dcop_66 Sandia/adder_dcop_67 Sandia/adder_dcop_68 Sandia/adder_dcop_69 Sandia/adder_trans_01 Sandia/adder_trans_02 Sandia/fpga_dcop_01 Sandia/fpga_dcop_02 Sandia/fpga_dcop_03 Sandia/fpga_dcop_04 Sandia/fpga_dcop_05 Sandia/fpga_dcop_06 Sandia/fpga_dcop_07 Sandia/fpga_dcop_08 Sandia/fpga_dcop_09 Sandia/fpga_dcop_10 Sandia/fpga_dcop_11 Sandia/fpga_dcop_12 Sandia/fpga_dcop_13 Sandia/fpga_dcop_14 Sandia/fpga_dcop_15 Sandia/fpga_dcop_16 Sandia/fpga_dcop_17 Sandia/fpga_dcop_18 Sandia/fpga_dcop_19 Sandia/fpga_dcop_20 Sandia/fpga_dcop_21 Sandia/fpga_dcop_22 Sandia/fpga_dcop_23 Sandia/fpga_dcop_24 Sandia/fpga_dcop_25 Sandia/fpga_dcop_26 Sandia/fpga_dcop_27 Sandia/fpga_dcop_28 Sandia/fpga_dcop_29 Sandia/fpga_dcop_30 Sandia/fpga_dcop_31 Sandia/fpga_dcop_32 Sandia/fpga_dcop_33 Sandia/fpga_dcop_34 Sandia/fpga_dcop_35 Sandia/fpga_dcop_36 Sandia/fpga_dcop_37 Sandia/fpga_dcop_38 Sandia/fpga_dcop_39 Sandia/fpga_dcop_40 Sandia/fpga_dcop_41 Sandia/fpga_dcop_42 Sandia/fpga_dcop_43 Sandia/fpga_dcop_44 Sandia/fpga_dcop_45 Sandia/fpga_dcop_46 Sandia/fpga_dcop_47 Sandia/fpga_dcop_48 Sandia/fpga_dcop_49 Sandia/fpga_dcop_50 Sandia/fpga_dcop_51 Sandia/fpga_trans_01 Sandia/fpga_trans_02 Sandia/init_adder1 Sandia/mult_dcop_01 Sandia/mult_dcop_02 Sandia/mult_dcop_03 Sandia/oscil_dcop_01 Sandia/oscil_dcop_02 Sandia/oscil_dcop_03 Sandia/oscil_dcop_04 Sandia/oscil_dcop_05 Sandia/oscil_dcop_06 Sandia/oscil_dcop_07 Sandia/oscil_dcop_08 Sandia/oscil_dcop_09 Sandia/oscil_dcop_10 Sandia/oscil_dcop_11 Sandia/oscil_dcop_12 Sandia/oscil_dcop_13 Sandia/oscil_dcop_14 Sandia/oscil_dcop_15 Sandia/oscil_dcop_16 Sandia/oscil_dcop_17 Sandia/oscil_dcop_18 Sandia/oscil_dcop_19 Sandia/oscil_dcop_20 Sandia/oscil_dcop_21 Sandia/oscil_dcop_22 Sandia/oscil_dcop_23 Sandia/oscil_dcop_24 Sandia/oscil_dcop_25 Sandia/oscil_dcop_26 Sandia/oscil_dcop_27 Sandia/oscil_dcop_28 Sandia/oscil_dcop_29 Sandia/oscil_dcop_30 Sandia/oscil_dcop_31 Sandia/oscil_dcop_32 Sandia/oscil_dcop_33 Sandia/oscil_dcop_34 Sandia/oscil_dcop_35 Sandia/oscil_dcop_36 Sandia/oscil_dcop_37 Sandia/oscil_dcop_38 Sandia/oscil_dcop_39 Sandia/oscil_dcop_40 Sandia/oscil_dcop_41 Sandia/oscil_dcop_42 Sandia/oscil_dcop_43 Sandia/oscil_dcop_44 Sandia/oscil_dcop_45 Sandia/oscil_dcop_46 Sandia/oscil_dcop_47 Sandia/oscil_dcop_48 Sandia/oscil_dcop_49 Sandia/oscil_dcop_50 Sandia/oscil_dcop_51 Sandia/oscil_dcop_52 Sandia/oscil_dcop_53 Sandia/oscil_dcop_54 Sandia/oscil_dcop_55 Sandia/oscil_dcop_56 Sandia/oscil_dcop_57 Sandia/oscil_trans_01 April 2003 * added matrices 917 to 922, from Michael Hohn Hohn/fd12 Hohn/fd15 Hohn/fd18 Hohn/sinc12 Hohn/sinc15 Hohn/sinc18 Jan 2003 * added matrices 865 to 916, from Alex Pothen, Stuart Norris, Kai Shen, Victor Pereyra, Alexander van Heukelum. In addition, many of the Harwell/Boeing matrices were found to have columns with unsorted row indices (the specification of that format states they must be sorted). Matrices in the Bai, HB (psmigr3), LPnetlib, and Simon directories were updated. Pothen/barth Pothen/barth4 Pothen/barth5 Pothen/bodyy4 Pothen/bodyy5 Pothen/bodyy6 Pothen/commanche_dual Pothen/mesh1e1 Pothen/mesh1em1 Pothen/mesh1em6 Pothen/mesh2e1 Pothen/mesh2em5 Pothen/mesh3e1 Pothen/mesh3em5 Pothen/onera_dual Pothen/pwt Pothen/shuttle_eddy Pothen/skirt Pothen/sphere2 Pothen/sphere3 Pothen/tandem_dual Pothen/tandem_vtx Norris/fv1 Norris/fv2 Norris/fv3 Norris/heart1 Norris/heart2 Norris/heart3 Norris/lung1 Norris/lung2 Norris/stomach Norris/torso1 Norris/torso2 Norris/torso3 Shen/e40r0100 Shen/shermanACa Shen/shermanACb Shen/shermanACd Pereyra/landmark vanHeukelum/cage3 vanHeukelum/cage4 vanHeukelum/cage5 vanHeukelum/cage6 vanHeukelum/cage7 vanHeukelum/cage8 vanHeukelum/cage9 vanHeukelum/cage10 vanHeukelum/cage11 vanHeukelum/cage12 vanHeukelum/cage13 vanHeukelum/cage14 vanHeukelum/cage15 Dec 2002 * added matrices with id's 842 to 864, from Andrew Cunnigham, Boeing, Pu Chen, The MathWorks, Lars Langemyr, and Wonsik Kim Cunningham/k3plates Cunningham/m3plates Cunningham/qa8fk Cunningham/qa8fm Boeing/bcsstk39 Chen/pkustk01 Chen/pkustk02 Chen/pkustk03 Chen/pkustk04 Chen/pkustk05 Chen/pkustk06 Chen/pkustk07 Chen/pkustk08 Chen/pkustk09 Chen/pkustk10 Chen/pkustk11 Chen/pkustk12 Chen/pkustk13 Chen/pkustk14 MathWorks/pivtol Kim/kim1 Kim/kim2 Langemyr/comsol March 2002 * MATLAB versions of Harwell/Boeing matrices created, and UFget added (interface to download the matrices into MATLAB). Matrix ID's defined. The collection consists of 841 matrices. Dec 2001 * matrices from Wim Bomhof Bomhof/circuit_1 Bomhof/circuit_2 Bomhof/circuit_3 Bomhof/circuit_4 Nov 2001 * matrices from Peter Hollinger, Daniel Okunbor, and Li Zhao Hollinger/g7jac010 Hollinger/g7jac010sc Hollinger/g7jac020 Hollinger/g7jac020sc Hollinger/g7jac040 Hollinger/g7jac040sc Hollinger/g7jac050sc Hollinger/g7jac060 Hollinger/g7jac060sc Hollinger/g7jac080 Hollinger/g7jac080sc Hollinger/g7jac100 Hollinger/g7jac100sc Hollinger/g7jac120 Hollinger/g7jac120sc Hollinger/g7jac140 Hollinger/g7jac140sc Hollinger/g7jac160 Hollinger/g7jac160sc Hollinger/g7jac180 Hollinger/g7jac180sc Hollinger/g7jac200 Hollinger/g7jac200sc Hollinger/jan99jac020 Hollinger/jan99jac020sc Hollinger/jan99jac040 Hollinger/jan99jac040sc Hollinger/jan99jac060 Hollinger/jan99jac060sc Hollinger/jan99jac080 Hollinger/jan99jac080sc Hollinger/jan99jac100 Hollinger/jan99jac100sc Hollinger/jan99jac120 Hollinger/jan99jac120sc Hollinger/mark3jac020 Hollinger/mark3jac020sc Hollinger/mark3jac040 Hollinger/mark3jac040sc Hollinger/mark3jac060 Hollinger/mark3jac060sc Hollinger/mark3jac080 Hollinger/mark3jac080sc Hollinger/mark3jac100 Hollinger/mark3jac100sc Hollinger/mark3jac120 Hollinger/mark3jac120sc Hollinger/mark3jac140 Hollinger/mark3jac140sc Okunbor/aft01 Okunbor/aft02 Zhao/Zhao1 Zhao/Zhao2 Oct 2001 * matrices from Klaus Gaertner Gaertner/big Gaertner/nopoly Gaertner/pesa July 2001 * matrices from Steve Hamm, Motorola Hamm/bcircuit Hamm/hcircuit Hamm/scircuit March 2001 * matrices from David Ronis Ronis/xenon1 Ronis/xenon2 Jan 2001 * matrices from David Long and Bob Melville, AT & T ATandT/onetone1 ATandT/onetone2 ATandT/pre2 ATandT/twotone Nov 2000 * matrix from Noel Brunetiere Brunetiere/thermal Nov 2000 * matrices from Karoly Nemeth Nemeth/nemeth01 Nemeth/nemeth02 Nemeth/nemeth03 Nemeth/nemeth04 Nemeth/nemeth05 Nemeth/nemeth06 Nemeth/nemeth07 Nemeth/nemeth08 Nemeth/nemeth09 Nemeth/nemeth10 Nemeth/nemeth11 Nemeth/nemeth12 Nemeth/nemeth13 Nemeth/nemeth14 Nemeth/nemeth15 Nemeth/nemeth16 Nemeth/nemeth17 Nemeth/nemeth18 Nemeth/nemeth19 Nemeth/nemeth20 Nemeth/nemeth21 Nemeth/nemeth22 Nemeth/nemeth23 Nemeth/nemeth24 Nemeth/nemeth25 Nemeth/nemeth26 Aug 1999 * random test matrices from the Gset collection Gset/G1 Gset/G10 Gset/G11 Gset/G12 Gset/G13 Gset/G14 Gset/G15 Gset/G16 Gset/G17 Gset/G18 Gset/G19 Gset/G2 Gset/G20 Gset/G21 Gset/G22 Gset/G23 Gset/G24 Gset/G25 Gset/G26 Gset/G27 Gset/G28 Gset/G29 Gset/G3 Gset/G30 Gset/G31 Gset/G32 Gset/G33 Gset/G34 Gset/G35 Gset/G36 Gset/G37 Gset/G38 Gset/G39 Gset/G4 Gset/G40 Gset/G41 Gset/G42 Gset/G43 Gset/G44 Gset/G45 Gset/G46 Gset/G47 Gset/G48 Gset/G49 Gset/G5 Gset/G50 Gset/G51 Gset/G52 Gset/G53 Gset/G54 Gset/G55 Gset/G56 Gset/G57 Gset/G58 Gset/G59 Gset/G6 Gset/G60 Gset/G61 Gset/G62 Gset/G63 Gset/G64 Gset/G65 Gset/G66 Gset/G67 Gset/G7 Gset/G8 Gset/G9 June 1999 * matrices from Ben Li Li/li Li/pli May 1998 * matrices from David Averous Averous/epb0 Averous/epb1 Averous/epb2 Averous/epb3 Feb 1998 * matrix from Dave Graham Graham/graham1 Dec 1997 * matrices from Andre Cote Cote/mplate Cote/vibrobox Nov 1997 * matrices from Jim Brethour Brethour/coater1 Brethour/coater2 Oct 1997 * matrices from Friedrich Grund Grund/b1_ss Grund/b2_ss Grund/b_dyn Grund/bayer01 Grund/bayer02 Grund/bayer03 Grund/bayer04 Grund/bayer05 Grund/bayer06 Grund/bayer07 Grund/bayer08 Grund/bayer09 Grund/bayer10 Grund/d_dyn Grund/d_dyn1 Grund/d_ss Grund/meg1 Grund/meg4 Grund/poli Grund/poli_large June 1997 * matrices from the NETLIB LP collection, converted to Harwell/Boeing format, Steve Bova, and Mauricio Resende (Qaplib problems). Bova/rma10 Qaplib/lp_nug05 Qaplib/lp_nug06 Qaplib/lp_nug07 Qaplib/lp_nug08 Qaplib/lp_nug12 Qaplib/lp_nug15 Qaplib/lp_nug20 Qaplib/lp_nug30 LPnetlib/lp_25fv47 LPnetlib/lp_80bau3b LPnetlib/lp_adlittle LPnetlib/lp_afiro LPnetlib/lp_agg LPnetlib/lp_agg2 LPnetlib/lp_agg3 LPnetlib/lp_bandm LPnetlib/lp_beaconfd LPnetlib/lp_blend LPnetlib/lp_bnl1 LPnetlib/lp_bnl2 LPnetlib/lp_bore3d LPnetlib/lp_brandy LPnetlib/lp_capri LPnetlib/lp_cre_a LPnetlib/lp_cre_b LPnetlib/lp_cre_c LPnetlib/lp_cre_d LPnetlib/lp_cycle LPnetlib/lp_czprob LPnetlib/lp_d2q06c LPnetlib/lp_d6cube LPnetlib/lp_degen2 LPnetlib/lp_degen3 LPnetlib/lp_dfl001 LPnetlib/lp_e226 LPnetlib/lp_etamacro LPnetlib/lp_fffff800 LPnetlib/lp_finnis LPnetlib/lp_fit1d LPnetlib/lp_fit1p LPnetlib/lp_fit2d LPnetlib/lp_fit2p LPnetlib/lp_ganges LPnetlib/lp_gfrd_pnc LPnetlib/lp_greenbea LPnetlib/lp_greenbeb LPnetlib/lp_grow15 LPnetlib/lp_grow22 LPnetlib/lp_grow7 LPnetlib/lp_israel LPnetlib/lp_kb2 LPnetlib/lp_ken_07 LPnetlib/lp_ken_11 LPnetlib/lp_ken_13 LPnetlib/lp_ken_18 LPnetlib/lp_lotfi LPnetlib/lp_maros LPnetlib/lp_maros_r7 LPnetlib/lp_modszk1 LPnetlib/lp_osa_07 LPnetlib/lp_osa_14 LPnetlib/lp_osa_30 LPnetlib/lp_osa_60 LPnetlib/lp_pds_02 LPnetlib/lp_pds_06 LPnetlib/lp_pds_10 LPnetlib/lp_pds_20 LPnetlib/lp_perold LPnetlib/lp_pilot LPnetlib/lp_pilot4 LPnetlib/lp_pilot87 LPnetlib/lp_pilot_ja LPnetlib/lp_pilot_we LPnetlib/lp_pilotnov LPnetlib/lp_qap12 LPnetlib/lp_qap15 LPnetlib/lp_qap8 LPnetlib/lp_recipe LPnetlib/lp_sc105 LPnetlib/lp_sc205 LPnetlib/lp_sc50a LPnetlib/lp_sc50b LPnetlib/lp_scagr25 LPnetlib/lp_scagr7 LPnetlib/lp_scfxm1 LPnetlib/lp_scfxm2 LPnetlib/lp_scfxm3 LPnetlib/lp_scorpion LPnetlib/lp_scrs8 LPnetlib/lp_scsd1 LPnetlib/lp_scsd6 LPnetlib/lp_scsd8 LPnetlib/lp_sctap1 LPnetlib/lp_sctap2 LPnetlib/lp_sctap3 LPnetlib/lp_share1b LPnetlib/lp_share2b LPnetlib/lp_shell LPnetlib/lp_ship04l LPnetlib/lp_ship04s LPnetlib/lp_ship08l LPnetlib/lp_ship08s LPnetlib/lp_ship12l LPnetlib/lp_ship12s LPnetlib/lp_sierra LPnetlib/lp_stair LPnetlib/lp_standata LPnetlib/lp_standgub LPnetlib/lp_standmps LPnetlib/lp_stocfor1 LPnetlib/lp_stocfor2 LPnetlib/lp_stocfor3 LPnetlib/lp_truss LPnetlib/lp_tuff LPnetlib/lp_vtp_base LPnetlib/lp_wood1p LPnetlib/lp_woodw LPnetlib/lpi_bgdbg1 LPnetlib/lpi_bgetam LPnetlib/lpi_bgindy LPnetlib/lpi_bgprtr LPnetlib/lpi_box1 LPnetlib/lpi_ceria3d LPnetlib/lpi_chemcom LPnetlib/lpi_cplex1 LPnetlib/lpi_cplex2 LPnetlib/lpi_ex72a LPnetlib/lpi_ex73a LPnetlib/lpi_forest6 LPnetlib/lpi_galenet LPnetlib/lpi_gosh LPnetlib/lpi_gran LPnetlib/lpi_greenbea LPnetlib/lpi_itest2 LPnetlib/lpi_itest6 LPnetlib/lpi_klein1 LPnetlib/lpi_klein2 LPnetlib/lpi_klein3 LPnetlib/lpi_mondou2 LPnetlib/lpi_pang LPnetlib/lpi_pilot4i LPnetlib/lpi_qual LPnetlib/lpi_reactor LPnetlib/lpi_refinery LPnetlib/lpi_vol1 LPnetlib/lpi_woodinfe May 1997 * matrices from Ed Rothberg, J. Mallya, and Mulvey (collected by Rothberg) Rothberg/3dtube Rothberg/cfd1 Rothberg/cfd2 Rothberg/gearbox Rothberg/struct3 Rothberg/struct4 Mallya/lhr04c Mallya/lhr07c Mallya/lhr10c Mallya/lhr11c Mallya/lhr14c Mallya/lhr17c Mallya/lhr34c Mallya/lhr71c Mulvey/finan512 Mulvey/pfinan512 March 1997 * matrices from Ansul Gupta, IBM Gupta/gupta1 Gupta/gupta2 Gupta/gupta3 Feb 1997 * more matrices from Steve Zitney, now at AspenTech Zitney/extr1 Zitney/hydr1 Zitney/radfr1 July 1996 * additional matrices from Yousef Saad TOKAMAK/utm1700b TOKAMAK/utm300 TOKAMAK/utm3060 TOKAMAK/utm5940 June 1996 * matrices from Bai Bai/af23560 Bai/bfwa398 Bai/bfwa62 Bai/bfwa782 Bai/bfwb398 Bai/bfwb62 Bai/bfwb782 Bai/bwm200 Bai/bwm2000 Bai/cdde1 Bai/cdde2 Bai/cdde3 Bai/cdde4 Bai/cdde5 Bai/cdde6 Bai/ck104 Bai/ck400 Bai/ck656 Bai/dw1024 Bai/dw256A Bai/dw256B Bai/dw4096 Bai/lop163 Bai/mhda416 Bai/mhdb416 Bai/odepa400 Bai/odepb400 Bai/olm100 Bai/olm1000 Bai/olm2000 Bai/olm500 Bai/olm5000 Bai/pde225 Bai/pde2961 Bai/pde900 Bai/qc2534 Bai/qc324 Bai/qh882 Bai/rbsa480 Bai/rbsb480 Bai/rdb2048 Bai/rdb5000 Bai/rdb968 Bai/rw136 Bai/rw496 Bai/rw5151 Bai/tub100 Bai/tub1000 May 1996 * matrices from Yousef Saad DRIVCAV/cavity01 DRIVCAV/cavity02 DRIVCAV/cavity03 DRIVCAV/cavity04 DRIVCAV/cavity05 DRIVCAV/cavity06 DRIVCAV/cavity07 DRIVCAV/cavity08 DRIVCAV/cavity09 DRIVCAV/cavity10 DRIVCAV/cavity11 DRIVCAV/cavity12 DRIVCAV/cavity13 DRIVCAV/cavity14 DRIVCAV/cavity15 DRIVCAV/cavity16 DRIVCAV/cavity17 DRIVCAV/cavity18 DRIVCAV/cavity19 DRIVCAV/cavity20 DRIVCAV/cavity21 DRIVCAV/cavity22 DRIVCAV/cavity23 DRIVCAV/cavity24 DRIVCAV/cavity25 DRIVCAV/cavity26 FIDAP/ex1 FIDAP/ex10 FIDAP/ex10hs FIDAP/ex11 FIDAP/ex12 FIDAP/ex13 FIDAP/ex14 FIDAP/ex15 FIDAP/ex18 FIDAP/ex19 FIDAP/ex2 FIDAP/ex20 FIDAP/ex21 FIDAP/ex22 FIDAP/ex23 FIDAP/ex24 FIDAP/ex25 FIDAP/ex26 FIDAP/ex27 FIDAP/ex28 FIDAP/ex29 FIDAP/ex3 FIDAP/ex31 FIDAP/ex32 FIDAP/ex33 FIDAP/ex35 FIDAP/ex36 FIDAP/ex37 FIDAP/ex4 FIDAP/ex40 FIDAP/ex5 FIDAP/ex6 FIDAP/ex7 FIDAP/ex8 FIDAP/ex9 May 1996 * matrices from Andre Garon Garon/garon1 Garon/garon2 May 1995 * additional matrices from Boeing, and new matrices from Ralph Goodwin Goodwin/goodwin Goodwin/rim Boeing/pcrystk02 Boeing/pcrystk03 Boeing/pct20stif June 1995 * matrices from NASA, collected by Alex Pothen These include xyz coordinates. Nasa/barth Nasa/barth4 Nasa/barth5 Nasa/nasa1824 Nasa/nasa2146 Nasa/nasa2910 Nasa/nasa4704 Nasa/nasasrb Nasa/pwt Nasa/shuttle_eddy Nasa/skirt June 1995 * added matrices from Boeing Boeing/bcsstk34 Boeing/bcsstk35 Boeing/bcsstk36 Boeing/bcsstk37 Boeing/bcsstk38 Boeing/bcsstm34 Boeing/bcsstm35 Boeing/bcsstm36 Boeing/bcsstm37 Boeing/bcsstm38 Boeing/bcsstm39 Boeing/crystk01 Boeing/crystk02 Boeing/crystk03 Boeing/crystm01 Boeing/crystm02 Boeing/crystm03 Boeing/ct20stif Boeing/msc00726 Boeing/msc01050 Boeing/msc01440 Boeing/msc04515 Boeing/msc10848 Boeing/msc23052 Boeing/nasa1824 Boeing/nasa2910 Boeing/nasa4704 Boeing/pwtk 1994 (approximate) * matrix from Steve Vavasis Vavasis/av41092 * matrices from Wei Shyy Shyy/shyy161 Shyy/shyy41 Aug. 1994 * from Song Wang Wang/swang1 Wang/swang2 Wang/wang1 Wang/wang2 Wang/wang3 Wang/wang4 July 1994 * added matrices from J. Mallya Mallya/lhr01 Mallya/lhr02 Mallya/lhr04 Mallya/lhr07 Mallya/lhr10 Mallya/lhr11 Mallya/lhr14 Mallya/lhr17 Mallya/lhr34 Mallya/lhr71 Jan 1994 * added matrices from Horst Simon Simon/appu Simon/bbmat Simon/venkat01 Simon/venkat25 Simon/venkat50 Simon/olafu Simon/raefsky1 Simon/raefsky2 Simon/raefsky3 Simon/raefsky4 Simon/raefsky5 Simon/raefsky6 Jan 1992 * added matrices from Steve Zitney, Cray Research Zitney/rdist1 Zitney/rdist2 Zitney/rdist3a July 1991 * matrices from Steve Hamm, Motorola Hamm/add20 Hamm/add32 Hamm/memplus 1991 * The UF sparse collection started as an ftp directory with just the 282 HB (Harwell/Boeing) matrices, in 1991. These matrices date well before 1991, of course. What is recorded in this log is when the matrices were added to this collection, not when they were first generated. HB/1138_bus HB/494_bus HB/662_bus HB/685_bus HB/abb313 HB/arc130 HB/ash219 HB/ash292 HB/ash331 HB/ash608 HB/ash85 HB/ash958 HB/bcspwr01 HB/bcspwr02 HB/bcspwr03 HB/bcspwr04 HB/bcspwr05 HB/bcspwr06 HB/bcspwr07 HB/bcspwr08 HB/bcspwr09 HB/bcspwr10 HB/bcsstk01 HB/bcsstk02 HB/bcsstk03 HB/bcsstk04 HB/bcsstk05 HB/bcsstk06 HB/bcsstk07 HB/bcsstk08 HB/bcsstk09 HB/bcsstk10 HB/bcsstk11 HB/bcsstk12 HB/bcsstk13 HB/bcsstk14 HB/bcsstk15 HB/bcsstk16 HB/bcsstk17 HB/bcsstk18 HB/bcsstk19 HB/bcsstk20 HB/bcsstk21 HB/bcsstk22 HB/bcsstk23 HB/bcsstk24 HB/bcsstk25 HB/bcsstk26 HB/bcsstk27 HB/bcsstk28 HB/bcsstk29 HB/bcsstk30 HB/bcsstk31 HB/bcsstk32 HB/bcsstk33 HB/bcsstm01 HB/bcsstm02 HB/bcsstm03 HB/bcsstm04 HB/bcsstm05 HB/bcsstm06 HB/bcsstm07 HB/bcsstm08 HB/bcsstm09 HB/bcsstm10 HB/bcsstm11 HB/bcsstm12 HB/bcsstm13 HB/bcsstm19 HB/bcsstm20 HB/bcsstm21 HB/bcsstm22 HB/bcsstm23 HB/bcsstm24 HB/bcsstm25 HB/bcsstm26 HB/bcsstm27 HB/beacxc HB/beaflw HB/beause HB/blckhole HB/bp_0 HB/bp_1000 HB/bp_1200 HB/bp_1400 HB/bp_1600 HB/bp_200 HB/bp_400 HB/bp_600 HB/bp_800 HB/can_1054 HB/can_1072 HB/can_144 HB/can_161 HB/can_187 HB/can_229 HB/can_24 HB/can_256 HB/can_268 HB/can_292 HB/can_445 HB/can_61 HB/can_62 HB/can_634 HB/can_715 HB/can_73 HB/can_838 HB/can_96 HB/curtis54 HB/dwt_1005 HB/dwt_1007 HB/dwt_1242 HB/dwt_162 HB/dwt_193 HB/dwt_198 HB/dwt_209 HB/dwt_221 HB/dwt_234 HB/dwt_245 HB/dwt_2680 HB/dwt_307 HB/dwt_310 HB/dwt_346 HB/dwt_361 HB/dwt_419 HB/dwt_492 HB/dwt_503 HB/dwt_512 HB/dwt_59 HB/dwt_592 HB/dwt_607 HB/dwt_66 HB/dwt_72 HB/dwt_758 HB/dwt_869 HB/dwt_87 HB/dwt_878 HB/dwt_918 HB/dwt_992 HB/eris1176 HB/fs_183_1 HB/fs_183_3 HB/fs_183_4 HB/fs_183_6 HB/fs_541_1 HB/fs_541_2 HB/fs_541_3 HB/fs_541_4 HB/fs_680_1 HB/fs_680_2 HB/fs_680_3 HB/fs_760_1 HB/fs_760_2 HB/fs_760_3 HB/gemat1 HB/gemat11 HB/gemat12 HB/gent113 HB/gr_30_30 HB/gre_1107 HB/gre_115 HB/gre_185 HB/gre_216a HB/gre_216b HB/gre_343 HB/gre_512 HB/hor_131 HB/ibm32 HB/illc1033 HB/illc1850 HB/impcol_a HB/impcol_b HB/impcol_c HB/impcol_d HB/impcol_e HB/jagmesh1 HB/jagmesh2 HB/jagmesh3 HB/jagmesh4 HB/jagmesh5 HB/jagmesh6 HB/jagmesh7 HB/jagmesh8 HB/jagmesh9 HB/jgl009 HB/jgl011 HB/jpwh_991 HB/lns_131 HB/lns_3937 HB/lns_511 HB/lnsp3937 HB/lnsp_131 HB/lnsp_511 HB/lshp1009 HB/lshp1270 HB/lshp1561 HB/lshp1882 HB/lshp2233 HB/lshp2614 HB/lshp3025 HB/lshp3466 HB/lshp_265 HB/lshp_406 HB/lshp_577 HB/lshp_778 HB/lund_a HB/lund_b HB/mahindas HB/mbeacxc HB/mbeaflw HB/mbeause HB/mcca HB/mcfe HB/nnc1374 HB/nnc261 HB/nnc666 HB/nos1 HB/nos2 HB/nos3 HB/nos4 HB/nos5 HB/nos6 HB/nos7 HB/orani678 HB/orsirr_1 HB/orsirr_2 HB/orsreg_1 HB/plat1919 HB/plat362 HB/plsk1919 HB/plskz362 HB/pores_1 HB/pores_2 HB/pores_3 HB/psmigr_1 HB/psmigr_2 HB/psmigr_3 HB/rgg010 HB/saylr1 HB/saylr3 HB/saylr4 HB/sherman1 HB/sherman2 HB/sherman3 HB/sherman4 HB/sherman5 HB/shl_0 HB/shl_200 HB/shl_400 HB/sstmodel HB/steam1 HB/steam2 HB/steam3 HB/str_0 HB/str_200 HB/str_400 HB/str_600 HB/watt_1 HB/watt_2 HB/well1033 HB/well1850 HB/west0067 HB/west0132 HB/west0156 HB/west0167 HB/west0381 HB/west0479 HB/west0497 HB/west0655 HB/west0989 HB/west1505 HB/west2021 HB/will199 HB/will57 HB/wm1 HB/wm2 HB/wm3 HB/young1c HB/young2c HB/young3c HB/young4c HB/zenios