Linear programming problems from H. Mittelmann 12 Jul 2006 ==================================== Benchmark of commercial LP solvers ==================================== H. Mittelmann (mittelmann@asu.edu) Logfiles of these runs at: http://plato.asu.edu/ftp/lp_logs/ This benchmark was run on a Linux-PC (3.2 GHz P4, 4 GB RDRAM, Linux-2.6). The MPS-datafiles for all testcases are in one of (see column "s") http://miplib.zib.de/ [1] http://plato.asu.edu/ftp/lptestset/ [2] http://www.sztaki.hu/~meszaros/public_ftp/lptestset/ (NETLIB[3], MISC[4], PROBLEMATIC[5], STOCHLP[6], KENNINGTON[7], INFEAS[8]) NOTE: files in [2-8] need to be expanded with emps in same directory! B/D/P: barrier/dual/primal simplex The following codes were tested: XP-B/D/P http://www.dashoptimization.com/ (XPRESS-MP-15.20) CPLEX-B/D/P http://www.cplex.com/ (ILOG-CPLEX-10.0) MOSEK-4.0.0.32 http://www.mosek.com (barrier) LOQO-6.06 http://www.princeton.edu/~rvdb/ LIPSOL linprog in Matlab 7.0 Times are user times in secs including input and crossover to a feasible basis for all codes except LOQO and LIPSOL. "$" without crossover. ===================================================== s problem CPLEX-B CPLEX-D/P MOSEK LOQO LIPSOL ===================================================== 2 cont1 952 2174/1530 5855 - - 2 cont11 1983 / 45727 - - 2 cont4 3505 2929/736 686 - - 2 cont1_l$ 3171 7102 2 cont11_l$ f 2081 1 dano3mip 16 36/23 14 85 14 4 dbic1 112 116/23 129 156 104 3 dfl001 14 20/46 12 152 13 2 fome12 169 141/407 57 485 88 2 fome13 87 669/1170 88 963 85 5 gen4 31 2/71 24 33 233 7 ken-18 11 12/68 21 274 25 5 l30 f 28/80 3 3 1691 4 lp22 6 40/72 7 62 9 4 mod2 10 82/192 20 61 25 2 neos 109 27/154 167 754 321 2 neos1 30 808/16 35 153 37 2 neos2 21 492/29 30 92 29 2 neos3 201 3208/5915 562 1505 502 4 nsct2 69 2/2 50 840 131 4 nug15 77 2912/1278 111 1855 118 2 nug20 1276 1490 26402 3092 2 nug08-3rd 1279 2367/ 1925 2004 2 pds-40 142 37/166 183 15908 252 2 pds-100 740 160/1586 1951 1673 3 qap12 10 203/112 15 201 17 3 qap15 74 2871/1230 107 1978 145 2 rail4284 239 5235/6476 386 4123 534 4 rlfprim 4 1/6 5 64 7 8 self 64 113/100 203 85 3194 2 sgpf5y6 11 3/2 18 fail 23 2 spal_004$ 4669 8020 m 4 stat96v1 390 188/641 47 4 stat96v2 >125000 29364/fail 390 4 stat96v4 8 396/553 30 6 storm-125 18 15/26 62 35 62 2 storm_1000 319 632/1482 916 1269 802 1 stp3d 168 1067/11715 196 2347 218 2 watson_2 49 192/346 66 313 70 4 world 10 94/237 25 44 28 ===================================================== Problem sizes problem rows columns nonzeros =============================================== cont1 160793 40398 399991 cont11 160793 80396 439989 cont4 160793 40398 398399 cont1_l 1918399 641598 5752001 cont11_l 1468599 981396 4403001 dano3mip_lp 3203 13873 79656 dbic1 43200 183235 1038761 dfl001 6072 12230 41873 fome12 24285 48920 167492 fome13 48569 97840 334984 gen4 1538 4297 110174 ken-18 105128 154699 512719 l30 2702 15380 64790 lp22 2959 13434 78994 mod2 35665 31728 220116 neos 479120 36786 1084461 neos1 131582 1892 468094 neos2 132569 1560 552596 neos3 512209 6624 1542816 nsct2 23004 14981 686396 nug15 6331 22275 110700 nug20 15240 72600 304800 nug08-3rd 19728 20448 139008 pds-40 66845 212859 605678 pds-100 156244 505360 1390539 qap12 3193 8856 44244 qap15 6331 22275 110700 rail4284 4284 1092610 12372358 rlfprim 57422 8052 264483 self 960 7364 1148845 sgpf5y6 246078 308634 902275 spal_004 10203 321696 46167908 stat96v1 5995 197472 588798 stat96v2 29089 957432 2852184 stat96v4 3173 62212 490472 stormG2-125 66186 157496 529317 stormG2_1000 528186 1259121 4228817 stp3d 159488 204880 662128 watson_2 352014 671861 1843716 world 35511 32734 220748 ===============================================