Hi Jack, A few comments.
- I notice that the code in that test used to select the best solver available had some errors, causing it to attempt to use MIPS (known to fail on this one) rather than Gurobi. I think I’ve fixed that problem in the latest version on the MOST GitHub repository. Try replacing your most/lib/t/t_most_w_ds.m file with this one <https://raw.githubusercontent.com/MATPOWER/most/master/lib/t/t_most_w_ds.m> and see if that works. - You may also have an issue with the installation of Gurobi on your machine. It seems odd to me that mpver does not print the version number, as it does for me. If typing mpver in a fresh Matlab session does not print the Gurobi version number, I’d be interested in following up with you to uncover whether there’s bug in MATPOWER or its an issue with your installation or what. - Finally, it’s always a challenge with numerical optimization to find the right options and tolerances to make tests pass across a bunch of different architectures, Matlab versions, solvers, etc. and this test is particularly troublesome, which is why I now moved it last. In the end, if everything else passes except this test, I wouldn’t worry about it. Ray > On Nov 19, 2018, at 5:31 AM, Jack Heron <[email protected]> wrote: > > > Hi, > > I have an update. This time I let ’test_most’ run for almost 45 minutes, and > it finally came up with the following warnings and error message. Does anyone > know what this could be? > > Any help would be much appreciated. > > Thanks, > > Jack > >>> test_most > t_most_3b_1_1_0........ok > t_most_3b_3_1_0........ok > t_most_30b_1_1_0.......ok > t_most_30b_3_1_0.......ok > t_most_fixed_res.......ok > t_most_30b_1_1_0_uc....ok > t_most_uc..............ok (156 of 260 skipped) > t_most_suc.............ok (111 of 185 skipped) > t_most_w_ds............> In mplinsolve (line 70) > In mips (line 467) > In qps_mips (line 180) > In qps_matpower (line 252) > In most (line 2029) > In t_most_w_ds (line 247) > In t_run_tests (line 61) > In test_most (line 57) > Warning: Matrix is close to singular or badly scaled. Results may be > inaccurate. RCOND = 1.450050e-16. >> In mplinsolve (line 70) > In mips (line 467) > In qps_mips (line 180) > In qps_matpower (line 252) > In most (line 2029) > In t_most_w_ds (line 247) > In t_run_tests (line 61) > In test_most (line 57) > Warning: Matrix is close to singular or badly scaled. Results may be > inaccurate. RCOND = 1.437052e-16. >> In mplinsolve (line 70) > In mips (line 467) > In qps_mips (line 180) > In qps_matpower (line 252) > In most (line 2029) > In t_most_w_ds (line 247) > In t_run_tests (line 61) > In test_most (line 57) > Warning: Matrix is close to singular or badly scaled. Results may be > inaccurate. RCOND = 1.169742e-19. >> In mplinsolve (line 70) > In mips (line 467) > In qps_mips (line 180) > In qps_matpower (line 252) > In most (line 2029) > In t_most_w_ds (line 247) > In t_run_tests (line 61) > In test_most (line 57) > Warning: Matrix is close to singular or badly scaled. Results may be > inaccurate. RCOND = 6.418324e-28. > Reference to non-existent field 'Z'. > > Error in t_most_w_ds (line 252) > t_is(mdo.results.Z, s.Z, 4, t); > > Error in t_run_tests (line 61) > feval( test_names{k}, ~verbose ); > > Error in test_most (line 57) > all_ok = t_run_tests( tests, verbose ); > > > >> On 16 Nov 2018, at 16:43, Jack Heron <[email protected]> wrote: >> >> Hi, >> >> I have just now installed matpower. When I run ‘test_matpower' everything is >> OK. However when I then run ‘test_most’, the operation gets stuck and seems >> to go on indefinitely - until I’m forced to terminate execution. >> >> I’ve tried several times, fresh install, etc. But it always gets stuck at >> the same place (’t_most_w_ds’). I have the latest version of Gurobi >> installed, and an older Matlab Optimisation toolbox. >> >> Can anyone help? >> >> This is the command window text I’m faced with: >> >>>> test_most >> t_most_3b_1_1_0........ok >> t_most_3b_3_1_0........ok >> t_most_30b_1_1_0.......ok >> t_most_30b_3_1_0.......ok >> t_most_fixed_res.......ok >> t_most_30b_1_1_0_uc....ok >> t_most_uc..............ok (156 of 260 skipped) >> t_most_suc.............ok (111 of 185 skipped) >> t_most_w_ds............Operation terminated by user during mplinsolve (line >> 70) >> >> >> In mips (line 467) >> dxdlam = mplinsolve([M dg; dg' sparse(neq, neq)], [-N; -g], opt.linsolver, >> mplinsolve_opt); >> >> In qps_mips (line 180) >> [x, f, eflag, output, lambda] = mips(p); >> >> In qps_matpower (line 252) >> qps_mips(H, c, A, l, u, xmin, xmax, x0, mips_opt); >> >> In most (line 2029) >> mdo.QP.lambda ] = qps_matpower( mdi.QP.H, mdi.QP.C, ... >> >> In t_most_w_ds (line 247) >> mdo = most(mdi, mpopt); >> >> In t_run_tests (line 61) >> feval( test_names{k}, ~verbose ); >> >> In test_most (line 57) >> all_ok = t_run_tests( tests, verbose ); >> >> >> ———— >> >> This is my output to ‘mpver’ : >> >>>> mpver >> >> MATPOWER Version 7.0b1 31-Oct-2018 >> MATLAB Version 9.1 25-Aug-2016 Release: (R2016b) >> Optimization Toolbox Version 7.5 25-Aug-2016 Release: (R2016b) >> MIPS Version 1.3 30-Oct-2018 >> MOST Version 1.0.1 30-Oct-2018 >> SDP_PF -- not installed -- >> YALMIP -- not installed -- >> BPMPD_MEX -- not installed -- >> CLP -- not installed -- >> CPLEX -- not installed -- >> GLPK -- not installed -- >> Gurobi Version >> IPOPT -- not installed -- >> KNITRO -- not installed -- >> MINOPF -- not installed -- >> MOSEK -- not installed -- >> PARDISO -- not installed -- >> PDIPMOPF -- not installed -- >> SCPDIPMOPF -- not installed -- >> SDPT3 -- not installed -- >> SeDuMi -- not installed -- >> TRALMOPF -- not installed -- >> Architecture: MACI64 >> >> MATPOWER 7.0b1 is distributed under the 3-clause BSD License. >> Please see the LICENSE file for details. >> >> ———— >> >> I’m running MacOS Mojave version 10.14 >> >> Thanks, >> >> Jack > > >
