|
From: | John W. Eaton |
Subject: | [Octave-bug-tracker] [bug #49904] Inversion of a Hermitian matrix produces incorrect result |
Date: | Wed, 21 Dec 2016 19:45:46 +0000 (UTC) |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Firefox/45.0 |
Update of bug #49904 (project octave): Item Group: Incorrect Result => Regression _______________________________________________________ Follow-up Comment #2: Here is what I see with 4.2.0: >> Rxx = [ 1.00003 + 0.00000i -0.68292 + 0.26406i -0.57670 + 0.25291i 0.34986 + 0.03335i -0.68292 - 0.26406i 0.94539 + 0.00000i 0.69006 - 0.03967i -0.43847 + 0.38619i -0.57670 - 0.25291i 0.69006 + 0.03967i 0.95260 + 0.00000i -0.34734 + 0.40728i 0.34986 - 0.03335i -0.43847 - 0.38619i -0.34734 - 0.40728i 0.98356 + 0.00000i] Rxx = 1.00003 + 0.00000i -0.68292 + 0.26406i -0.57670 + 0.25291i 0.34986 + 0.03335i -0.68292 - 0.26406i 0.94539 + 0.00000i 0.69006 - 0.03967i -0.43847 + 0.38619i -0.57670 - 0.25291i 0.69006 + 0.03967i 0.95260 + 0.00000i -0.34734 + 0.40728i 0.34986 - 0.03335i -0.43847 - 0.38619i -0.34734 - 0.40728i 0.98356 + 0.00000i >> Rxx * inv (Rxx) ans = 3.8447e+03 + 4.1605e+03i -4.2505e+02 - 5.3359e+03i 2.1636e+02 - 2.2698e+03i -2.2737e-13 - 1.1369e-12i 4.8212e+03 - 6.0472e+03i 4.0808e+03 + 6.5882e+03i 2.5055e+03 + 2.8447e+03i 0.0000e+00 + 1.3642e-12i 5.6085e+03 - 5.2994e+03i 4.1757e+03 + 6.4500e+03i 2.6433e+03 + 2.2534e+03i -6.8212e-13 + 0.0000e+00i -5.1015e+03 - 7.0805e+03i 1.2351e+03 - 1.1137e+04i 4.2633e-13 - 6.3810e+03i 1.0000e+00 - 2.2737e-13i >> inv (Rxx) ans = 7918.16 + 0.00i 4705.22 - 5155.62i 3582.48 - 1374.29i -2047.24 - 6383.13i 4705.22 - 5155.62i 6155.56 + 0.00i 3022.13 + 1516.30i 2940.40 - 5126.34i 3582.48 - 1374.29i 3022.13 + 1516.30i 1861.79 + 0.00i 181.66 - 3243.85i -2047.24 - 6383.13i 2940.40 - 5126.34i 181.66 - 3243.85i 5676.72 + 0.00i and with 4.0.3: >> Rxx = [ 1.00003 + 0.00000i -0.68292 + 0.26406i -0.57670 + 0.25291i 0.34986 + 0.03335i -0.68292 - 0.26406i 0.94539 + 0.00000i 0.69006 - 0.03967i -0.43847 + 0.38619i -0.57670 - 0.25291i 0.69006 + 0.03967i 0.95260 + 0.00000i -0.34734 + 0.40728i 0.34986 - 0.03335i -0.43847 - 0.38619i -0.34734 - 0.40728i 0.98356 + 0.00000i] Rxx = 1.00003 + 0.00000i -0.68292 + 0.26406i -0.57670 + 0.25291i 0.34986 + 0.03335i -0.68292 - 0.26406i 0.94539 + 0.00000i 0.69006 - 0.03967i -0.43847 + 0.38619i -0.57670 - 0.25291i 0.69006 + 0.03967i 0.95260 + 0.00000i -0.34734 + 0.40728i 0.34986 - 0.03335i -0.43847 - 0.38619i -0.34734 - 0.40728i 0.98356 + 0.00000i >> Rxx * inv (Rxx) ans = 1.00000 + 0.00000i -0.00000 - 0.00000i 0.00000 - 0.00000i -0.00000 - 0.00000i -0.00000 - 0.00000i 1.00000 + 0.00000i -0.00000 + 0.00000i -0.00000 + 0.00000i -0.00000 - 0.00000i -0.00000 + 0.00000i 1.00000 - 0.00000i -0.00000 + 0.00000i 0.00000 + 0.00000i 0.00000 + 0.00000i 0.00000 + 0.00000i 1.00000 - 0.00000i >> inv (Rxx) ans = 7918.16 + 0.00i 4705.22 - 5155.62i 3582.48 - 1374.29i -2047.24 - 6383.13i 4705.22 + 5155.62i 6155.56 + 0.00i 3022.13 + 1516.30i 2940.40 - 5126.34i 3582.48 + 1374.29i 3022.13 - 1516.30i 1861.79 + 0.00i 181.66 - 3243.85i -2047.24 + 6383.13i 2940.40 + 5126.34i 181.66 + 3243.85i 5676.72 + 0.00i Something seems messed up with the signs here. I'll take a look and see if I can spot what changed. _______________________________________________________ Reply to this item at: <http://savannah.gnu.org/bugs/?49904> _______________________________________________ Message sent via/by Savannah http://savannah.gnu.org/
[Prev in Thread] | Current Thread | [Next in Thread] |