1a13f5d7d809b5ba48fa6dc672d7ad1216b86063 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/31132a0a6ae4f1817fb5f92235091d224b92b823 31132a0a6ae4f1817fb5f92235091d224b92b823

2009-06-10T15:06:45-07:00

Update for latest PETSc changes, especially PCShell 62a89d1674298a5ce91d7c343b1ddc20db1c00d7 Jed Brown jed@59A2.org c4ec55acd7bd70d2e532247daba08041d99be99a Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/1a13f5d7d809b5ba48fa6dc672d7ad1216b86063 1a13f5d7d809b5ba48fa6dc672d7ad1216b86063

2008-12-09T05:37:34-08:00

Update for recent PETSc changes 74c40120ae85e0be7d448809082840f79340de58 Jed Brown jed@59A2.org 7b5a7515f45d5c28b08d2d20d77c9dd6e347d9d1 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/c4ec55acd7bd70d2e532247daba08041d99be99a c4ec55acd7bd70d2e532247daba08041d99be99a

2008-08-15T03:46:39-07:00

Add plain C support, testing conditioning of perturbed matrices Makefile: Allow a plain C version of PETSc to be used stokes.C: Testing conditioning of perturbed matrices Add option to filter matrix for testing effect of imposing Dirichlet conditions by zeroing diagonal. The option are only effective if -pc_type 1 (Q1 finite element) preconditioning is chosen. There are two new options: -zeroN Int number of rows and columns to eliminate -zeroV Scalar value to place on the diagonal We can test the conditioning of the velocity matrix with -vel_ksp_type preonly -vel_pc_type ksp -vel_ksp_ksp_rtol 1e-10 \ -vel_ksp_ksp_gmres_restart 500 -vel_ksp_pc_type none \ -vel_ksp_ksp_monitor_singular_value and the conditioning of the Schur complement with -schur_pc_type none -svel_ksp_type preonly -svel_pc_type ksp \ -svel_ksp_rtol 1e-10 -svel_ksp_pc_type ilu -schur_ksp_gmres_restart 500 \ -schur_ksp_monitor_singular_value Note that the divergence matrices are still applied to spectral order and matrix-free. Hence the procedure here is only a first-order approximation to the effect on conditioning since we are not removing the appropriate rows and columns from these matrix-free operators. 73b66f32c27df7282ecdb842e8ee1091e2d9ec01 Jed Brown jed@59A2.org c4838ef0bcca3fa9d05fb08120c674022eccf6fa Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/7b5a7515f45d5c28b08d2d20d77c9dd6e347d9d1 7b5a7515f45d5c28b08d2d20d77c9dd6e347d9d1

2008-07-17T05:37:07-07:00

Cleanup of unused variables, restrict chebyshev.[ch] to C89. c69718741fcbf80a1a494c8d08cb9284d24356b5 Jed Brown jed@59A2.org fcd7e8788b6fd9829379f021c0d61079f4c63960 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/c4838ef0bcca3fa9d05fb08120c674022eccf6fa c4838ef0bcca3fa9d05fb08120c674022eccf6fa

2008-05-30T08:30:53-07:00

Another cheap, but less robust preconditioner. d7537f04732701894e040c8489d4e543d809f0af Jed Brown jed@59A2.org 7f19bd6f1ee1e3ca783c13c9b8e826b30d8af4cc Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/fcd7e8788b6fd9829379f021c0d61079f4c63960 fcd7e8788b6fd9829379f021c0d61079f4c63960

2008-05-16T13:44:02-07:00

Added an upper triangular preconditioner and flexible matrix format. Select the upper triangular preconditioner with -pc_saddle_type 1. 5330e0cb29b93899752e04cfba287a600c3b2204 Jed Brown jed@59A2.org fcfe7a61da774f900cc5cefa2820944f9d0d812c Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/7f19bd6f1ee1e3ca783c13c9b8e826b30d8af4cc 7f19bd6f1ee1e3ca783c13c9b8e826b30d8af4cc

2008-04-30T07:25:31-07:00

Enable diagonal preconditioning of Schur complement using effective viscosity. 39b9aa94935cff2fb03294ca063fe61a905a51ff Jed Brown jed@59A2.org acd5393ac6a2eaa35e97692b3e295092f272fd9c Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/fcfe7a61da774f900cc5cefa2820944f9d0d812c fcfe7a61da774f900cc5cefa2820944f9d0d812c

2008-04-28T02:34:06-07:00

Minor cleanup, fleshed out README. e09dbf3864f335a14ada9a6965e7692773a86105 Jed Brown jed@59A2.org 6989498ef874e4b8af2d58de175e0a699329ec7b Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/acd5393ac6a2eaa35e97692b3e295092f272fd9c acd5393ac6a2eaa35e97692b3e295092f272fd9c

2008-04-28T01:57:15-07:00

Added README. 9de8a6b9a9050af433e033f1490e28682476666e Jed Brown jed@59A2.org 9c3184c8e5c46f0208e1bd5fbbbd3f3859dc73da Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/6989498ef874e4b8af2d58de175e0a699329ec7b 6989498ef874e4b8af2d58de175e0a699329ec7b

2008-04-28T01:35:52-07:00

Fix FFTW wisdom check, wrap CppAD stuff with preprocessor. 05836de7bab816d151bec055b7244a613f17122f Jed Brown jed@59A2.org c19e38fe77b13c32237cebc5694001c2991de372 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/9c3184c8e5c46f0208e1bd5fbbbd3f3859dc73da 9c3184c8e5c46f0208e1bd5fbbbd3f3859dc73da

2008-04-22T07:04:02-07:00

Started outflow boundary condition. cd044a17d1de923f2a0eee86cae6e22682cdf07d Jed Brown jed@59A2.org 3d76efd91a7d1903c5ab0cb6d147bf79003a340d Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/c19e38fe77b13c32237cebc5694001c2991de372 c19e38fe77b13c32237cebc5694001c2991de372

2008-04-12T13:37:36-07:00

Added parameters controlling mixed condition scaling. Another boundary. c7523ca9eee0d119b09c27091db7e287d0ec1a77 Jed Brown jed@59A2.org 7435421b2f73baf11da4488e6c3cbde0ac8a488a Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/3d76efd91a7d1903c5ab0cb6d147bf79003a340d 3d76efd91a7d1903c5ab0cb6d147bf79003a340d

2008-04-12T10:36:57-07:00

Minor cleanup. 9c97d60a76865987ea609052cc84655c57739c31 Jed Brown jed@59A2.org 0da905f9b37ff801d620dfa364e91044fefcf30a Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/7435421b2f73baf11da4488e6c3cbde0ac8a488a 7435421b2f73baf11da4488e6c3cbde0ac8a488a

2008-04-12T08:50:49-07:00

Implemented simple finite difference preconditioning using CppAD. ae37790540f7491c2e9e2752da55f85d019bc7a8 Jed Brown jed@59A2.org 4c9ff3db1ad754e09f34624e05347b3ea18fd687 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/0da905f9b37ff801d620dfa364e91044fefcf30a 0da905f9b37ff801d620dfa364e91044fefcf30a

2008-04-09T14:58:13-07:00

Implemented approximating the spectral matrix using a coloring. Unfortunately, this is even worse quality than the finite element approximation. 191e0e9ccc3de8c9b191d34b0c87a7fd560eedb7 Jed Brown jed@59A2.org 34b8d5f6b9b2b77a35b24452a4c81baa8c8a6e66 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/4c9ff3db1ad754e09f34624e05347b3ea18fd687 4c9ff3db1ad754e09f34624e05347b3ea18fd687

2008-04-09T11:25:23-07:00

Form the lumped global mass matrix. It seems slightly better, but FD still wins. 6cec66a9b64d97e7630e6827a0d2de6aa091c870 Jed Brown jed@59A2.org 93729ba6559476efdc7e334ad4a47275e2fe11a6 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/34b8d5f6b9b2b77a35b24452a4c81baa8c8a6e66 34b8d5f6b9b2b77a35b24452a4c81baa8c8a6e66

2008-04-09T10:57:15-07:00

Added formation of the element mass matrix. This still doesn't seem to help. The finite difference preconditioner is still significantly stronger. 74a45cb0b48563233a835788f215c20637c93c0b Jed Brown jed@59A2.org 02b61febc0cc8e1f1b9ccc04e81744bd4275d413 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/93729ba6559476efdc7e334ad4a47275e2fe11a6 93729ba6559476efdc7e334ad4a47275e2fe11a6

2008-04-09T10:32:40-07:00

Finite element preconditioner works, but is actually not as strong as the finite difference one. After much bug hunting, I became convinced that the inner part is actually correct, but the Q1 finite element preconditioner is just not as good as the finite difference preconditioner, at least applied naively. I believe the root of the problem is that we need to apply the mass matrix in order to get the scaling correct. 2ac0f4ad328b33926426c60bd210ada55bf69fc3 Jed Brown jed@59A2.org ebb61e91f0c5f8c2729565666202d47295f9eb7c Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/02b61febc0cc8e1f1b9ccc04e81744bd4275d413 02b61febc0cc8e1f1b9ccc04e81744bd4275d413

2008-04-08T12:28:51-07:00

Finite element discretization on CGL nodes. 6788e75dba3c35e766d895904c47b2cb850d26cd Jed Brown jed@59A2.org d0847b62e26fb0e8f98bd581a4ba9ed24d7a6569 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/ebb61e91f0c5f8c2729565666202d47295f9eb7c ebb61e91f0c5f8c2729565666202d47295f9eb7c

2008-04-07T13:54:07-07:00

Constrain extra degree of freedom for mixed condition. 1750dde910d6602eb39f5c71d85d9b18901df781 Jed Brown jed@59A2.org 65a6512a48532fe8b390684cfb59af93ffa51e9a Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/d0847b62e26fb0e8f98bd581a4ba9ed24d7a6569 d0847b62e26fb0e8f98bd581a4ba9ed24d7a6569

2008-04-05T06:02:51-07:00

Preconditioner works with mixed conditions. Convergence is slow. 506ab75e14ff3d787e02b572e64383a48e000561 Jed Brown jed@59A2.org f8e7ec15df805e2f16f028473ae378fda55db47a Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/65a6512a48532fe8b390684cfb59af93ffa51e9a 65a6512a48532fe8b390684cfb59af93ffa51e9a

2008-04-05T06:01:49-07:00

Get the normal direction correct for Chebyshev ordering. 85c26c059b2907599c849531e562e1113771ea6b Jed Brown jed@59A2.org 26a6464810c0e4b6aa440701bb937fed1ed10d66 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/f8e7ec15df805e2f16f028473ae378fda55db47a f8e7ec15df805e2f16f028473ae378fda55db47a

2008-04-04T08:47:21-07:00

Neumann handling, but not yet in preconditioner. 7db1eacab2def2da4d907538938ec40b0964a0a1 Jed Brown jed@59A2.org 049fe0287ba842cf891ad895ced5154ba08f5a19 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/26a6464810c0e4b6aa440701bb937fed1ed10d66 26a6464810c0e4b6aa440701bb937fed1ed10d66

2008-04-03T06:58:21-07:00

Also write force. Started Neumann boundary condition. e3a74521de15ab95cc2ac0e6a2d04e890cda39b0 Jed Brown jed@59A2.org 1581032719fb3140a83a856fe24945221ae4cc16 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/049fe0287ba842cf891ad895ced5154ba08f5a19 049fe0287ba842cf891ad895ced5154ba08f5a19

2008-04-03T00:12:24-07:00

2008-04-02T23:40:41-07:00

Power law fluid, continuation, simple VTK output. 621b9a0ea3c80c17966debff000c715c39408c1e Jed Brown jed@59A2.org f52f648f91acf51bf43c446c1fcda0206b5128bd Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/1581032719fb3140a83a856fe24945221ae4cc16 1581032719fb3140a83a856fe24945221ae4cc16

2008-04-03T00:12:24-07:00

2008-04-01T08:08:41-07:00

Precondition velocity system with a laplace-type operator. 44b5877b21b38e07887f5cae363bd03080b21048 Jed Brown jed@59A2.org 79f27f9c1e987941c1b6e040ab63f574081fbd5c Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/f52f648f91acf51bf43c446c1fcda0206b5128bd f52f648f91acf51bf43c446c1fcda0206b5128bd

2008-04-03T00:12:24-07:00

2008-04-01T06:15:59-07:00

Fixed Schur complement tampering, null space removal. ab8a021251e7783d02baf340aed4424a15c04fe7 Jed Brown jed@59A2.org ec6ba743f30b58d38325f6ac8e0d69a562f1d59b Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/79f27f9c1e987941c1b6e040ab63f574081fbd5c 79f27f9c1e987941c1b6e040ab63f574081fbd5c

2008-04-03T00:10:48-07:00

2008-04-01T06:15:11-07:00

Quieter build. 7b37b63a233ba58682afc9a17c21c77a2024c32c Jed Brown jed@59A2.org fc9205b67efc9f50bde286c1ad108edbe7de9cc0 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/ec6ba743f30b58d38325f6ac8e0d69a562f1d59b ec6ba743f30b58d38325f6ac8e0d69a562f1d59b

2008-04-01T01:14:36-07:00

Many fixes to decomposition and setting up domain. Only use symmetric strain. 8101014e8f45cd24d7863eb807e12feaedcd2cf4 Jed Brown jed@59A2.org 020b523942afa00c37423567706e0e62881a3e49 Jed Brown jed@59A2.org http://github.com/jedbrown/spectral-petsc/commit/fc9205b67efc9f50bde286c1ad108edbe7de9cc0 fc9205b67efc9f50bde286c1ad108edbe7de9cc0

2008-03-30T15:16:16-07:00

Much cleaner design and theory. 73f3c3b2522fa975dfd12320b617e754d6ed99db Jed Brown jed@59A2.org