changelog

来自mit的fdtd开放性源代码meep

    M ./src/fields.cpp r2
    M ./src/meep.hpp r2
    M ./src/sources.cpp -3 +7 r2
    M ./src/step.cpp -2 +2 r2
    M ./src/update_e_from_d.cpp r1

Thu Jun 19 18:16:54 EDT 2008  [e-mail]
  * send H, not B, when H==B

    M ./src/boundaries.cpp -15 +17

Thu Jun 19 18:04:45 EDT 2008  [e-mail]
  * in preparation for Dcyl, use function for cyclic shifts of directions

    M ./src/meep/vec.hpp +11
    M ./src/monitor.cpp -6 +2
    M ./src/step_b.cpp -4 +6
    M ./src/step_d.cpp -4 +5
    M ./src/structure.cpp -1 +1
    M ./src/update_e_from_d.cpp -8 +8
    M ./src/update_h_from_b.cpp -8 +8

Thu Jun 19 18:03:48 EDT 2008  [e-mail]
  * can't force_consistency for H, only for B

    M ./src/cw_fields.cpp -4 +4
    M ./src/step.cpp -2 +1

Thu Jun 19 17:41:08 EDT 2008  [e-mail]
  * handle u == NULL in update_EDHB

    M ./src/step_generic.cpp -1 +3

Thu Jun 19 15:58:57 EDT 2008  [e-mail]
  * use HUGE_VAL for infinity, since this is in c90 standard and in C++, and is equivalent to INFINITY on IEEE754 systems; don't define "infinity" twice in meep.scm

    M ./libctl/meep.scm.in -2
    M ./src/meep.hpp -5 +1

Thu Jun 19 15:49:12 EDT 2008  [e-mail]
  * implement output-mu and related functions to output/integrate Permeability

    M ./libctl/meep.scm.in +1
    M ./src/h5fields.cpp -4 +34
    M ./src/integrate.cpp -7 +37
    M ./src/loop_in_chunks.cpp +2
    M ./src/monitor.cpp -1 +4

Thu Jun 19 15:48:50 EDT 2008  [e-mail]
  * whoops, typo (set_mu wrote to inveps instead of invmu)

    M ./src/anisotropic_averaging.cpp -2 +2

Thu Jun 19 02:30:32 EDT 2008  [e-mail]
  * fix a couple of memory leaks

    M ./src/boundaries.cpp +1
    M ./src/dft.cpp +2

Thu Jun 19 01:17:18 EDT 2008  [e-mail]
  * only allocate separate B and H fields if necessary + many related changed
  
  now, unnecessary H fields are set == B at the pointer level; H is 
  allocated lazily, just before needed, in PML regions or when mu!=1,
  to guarantee consistency
  
  allow invmu to be NULL
  allow user-specification of mu != 1 (although no averaging is done yet)
  
  don't use PML equations if unnecessary - in the PML regions, some 
  components may still be updated with the ordinary equations
  
  fix potential bug in update_EDHB where the PML equations might not have 
  been used if g1 and g2 were swapped, since that also swaps sigsize_dsig1 
  
  the PML equations in nonlinear materials now use the fields, not the 
  fields multiplied by (1+sigma) etc., to determine the nonlinear 
  coefficient, which seems more correct (and allows us to switch to the 
  non-PML equations in more cases)
  

    M ./src/anisotropic_averaging.cpp -1 +7
    M ./src/boundaries.cpp -2 +21
    M ./src/energy_and_flux.cpp -1 +3
    M ./src/fields.cpp -1 +10
    M ./src/meep.hpp -1 +1
    M ./src/meep/mympi.hpp +2
    M ./src/mympi.cpp +14
    M ./src/step_generic.cpp -19 +29
    M ./src/structure.cpp -1 +1
    M ./src/update_h_from_b.cpp -19 +38

Thu Jun 19 01:12:53 EDT 2008  [e-mail]
  * in mix_with, handle case where two structures don't have the same inveps/invmu/conductivity allocated

    M ./src/structure.cpp -10 +47

Thu Jun 19 01:04:26 EDT 2008  [e-mail]
  * updated numbers in harmonics test (slightly changed by new PML)

    M ./tests/harmonics.cpp -4 +4

Thu Jun 19 00:12:11 EDT 2008  [e-mail]
  * silence compiler warnings

    M ./src/energy_and_flux.cpp -1 +1
    M ./src/integrate.cpp +1

Wed Jun 18 22:14:24 EDT 2008  [e-mail]
  * silence some compiler warnings

    M ./src/h5file.cpp -4 +3

Wed Jun 18 21:48:59 EDT 2008  [e-mail]
  * have_component should return true if *any* of the owned chunks have the component, not just the first owned chunk -- this doesn't seem to have caused problems in the past, but it 's a bit unsafe

    M ./src/fields.cpp -2 +2

Wed Jun 18 18:00:46 EDT 2008  [e-mail]
  * compute magnetic energy as B*H, not |H|^2

    M ./src/energy_and_flux.cpp -2 +3

Wed Jun 18 17:18:50 EDT 2008  [e-mail]
  * don't use conductivity and chi2/chi3 in chunks where they are zero

    M ./src/anisotropic_averaging.cpp -30
    M ./src/structure.cpp -4 +63

Wed Jun 18 17:16:51 EDT 2008  [e-mail]
  * bug fix in new PML: f_prev should be D-P, not D

    M ./src/energy_and_flux.cpp +3
    M ./src/step_d.cpp -5 +7

Wed Jun 18 01:17:37 EDT 2008  [e-mail]
  * added conductivity & mu support to libctl (although   mu support still needs to be enabled in C++); conductivity seems to work in simple tests

    M ./libctl/meep.scm.in -1 +10
    M ./libctl/structure.cpp -69 +142
    M ./src/anisotropic_averaging.cpp -20 +19
    M ./src/meep.hpp -3 +7
    M ./src/structure.cpp +3

Wed Jun 18 00:45:35 EDT 2008  [e-mail]
  * fix (harmless) compiler warnings

    M ./src/h5fields.cpp -4 +6

Tue Jun 17 23:43:39 EDT 2008  [e-mail]
  * initial stab at general conductivity support (diagonal anisotropic B or D conductors) - untested (but compiles and shouldn't affect programs not using the conducitivity)

    M ./src/anisotropic_averaging.cpp -1 +31
    M ./src/meep.hpp -21 +12
    M ./src/meep_internals.hpp +24
    M ./src/step.cpp +3
    M ./src/step_b.cpp -5 +10
    M ./src/step_d.cpp -5 +10
    M ./src/step_generic.cpp -19 +63
    M ./src/structure.cpp -27 +83

Tue Jun 17 23:42:49 EDT 2008  [e-mail]
  * error thresholds were too low (passed on x86 in extended-precision mode, but not on x86_64 with ordinary double precision); also, work around weird gcc problem with array of fields/structures

    M ./tests/symmetry.cpp -23 +27

Tue Jun 17 23:41:47 EDT 2008  [e-mail]
  * don't forget to call set_volume in set_mu

    M ./src/anisotropic_averaging.cpp -1 +4

Tue Jun 17 23:41:05 EDT 2008  [e-mail]
  * work around mysterious valgrind complaint about uninitialized var when running bragg_transmission.dac; also, slight code cleanup (PASTE macro was unnecessary)

    M ./src/step_generic.cpp -12 +22

Tue Jun 17 18:04:53 EDT 2008  [e-mail]
  * slight clarification (behavior unchanged)

    M ./src/structure.cpp -1 +1

Tue Jun 17 17:22:39 EDT 2008  [e-mail]
  * remove redundnant test

    M ./src/step_generic.cpp -12 +6

Tue Jun 17 17:05:44 EDT 2008  [e-mail]
  * use macros to reduce cut-and-paste in update_EDHB, and fix several bugs in index computations

    M ./src/step_generic.cpp -172 +82

Tue Jun 17 16:52:16 EDT 2008  [e-mail]
  * avoid high-symmetry monitor point

    M ./tests/2D_convergence.cpp -1 +1

Tue Jun 17 15:49:34 EDT 2008  [e-mail]
  * remove unused variables/computations

    M ./src/step_generic.cpp -8

Tue Jun 17 15:18:54 EDT 2008  [e-mail]
  * added g2 && !g1 case for non-PML nonlinear

    M ./src/step_generic.cpp +8

Tue Jun 17 14:43:25 EDT 2008  [e-mail]
  * added two more missing files

    A ./src/step_b.cpp
    A ./src/update_h_from_b.cpp

Tue Jun 17 14:42:57 EDT 2008  [e-mail]
  * added step_generic

    A ./src/step_generic.cpp

Tue Jun 17 11:50:14 EDT 2008  [e-mail]
  * set default Rabs to 1e-15 (todo: make this user-configurable, along with PML profile)

    M ./src/structure.cpp -1 +1

Tue Jun 17 11:49:50 EDT 2008  [e-mail]
  * handle off-diagonal inveps

    M ./src/update_e_from_d.cpp -2 +2

Tue Jun 17 11:39:53 EDT 2008  [e-mail]
  * optimization - only call average_with_backup if is_mine

    M ./src/energy_and_flux.cpp -4 +5

Sun Jun 15 22:09:54 EDT 2008  [e-mail]
  * lower tolerance still works

    M ./tests/two_dimensional.cpp -1 +1

Thu Jun 12 20:15:18 EDT 2008  [e-mail]
  * tolerances were too low (the energy numbers fluctuate a little from timestep to timestep, and the old tolerance was right on the borderline)

    M ./tests/two_dimensional.cpp -1 +1

Thu Jun 12 19:02:24 EDT 2008  [e-mail]
  * resolve conflicting patch

    M ./src/update_e_from_d.cpp -60 +12

Thu Jun 12 18:33:34 EDT 2008  [e-mail]
  * separate f_prev and f_backup, add synchronize_magnetic_fields function

    M ./src/cw_fields.cpp -2 +2
    M ./src/energy_and_flux.cpp -53 +46
    M ./src/fields.cpp +4
    M ./src/meep.hpp -6 +9
    M ./src/step_d.cpp -1 +6

Thu Jun 12 15:23:41 EDT 2008  [e-mail]
  * forgot to record this

    M ./src/structure.cpp -18

Thu Jun 12 15:22:03 EDT 2008  [e-mail]
  * slight cleanup

    M ./src/meep.hpp -3 +1

Thu Jun 12 14:44:04 EDT 2008  [e-mail]
  * big patch to rewrite timestepping and PML, and add mu/B

    M ./src/Makefile.am -12 +3
    M ./src/anisotropic_averaging.cpp -1 +16
    M ./src/cw_fields.cpp -29 +24
    M ./src/energy_and_flux.cpp -35 +50
    M ./src/fields.cpp -43 +4
    M ./src/initialize.cpp +1
    M ./src/meep.hpp -15 +45
    M ./src/meep/vec.hpp -19 +24
    M ./src/step.cpp -8 +16
    M ./src/step_d.cpp -150 +40
    R ./src/step_h.cpp
    M ./src/structure.cpp -69 +93
    M! ./src/update_e_from_d.cpp -94 +14
    M ./src/vec.cpp -1 +7
    M ./tests/harmonics.cpp -7 +7
    M ./tests/symmetry.cpp -2

Thu Jun 12 20:07:07 EDT 2008  [e-mail]
  * fixed relative phase of E and H sources (which were wrong by half a timestep); also, tolerance in two_dimensional.cpp was too low (changing the runtime by half a timestep changed the energy by more than the tolerance for the decay rate)

    M ./src/sources.cpp +3
    M ./src/step.cpp -2 +4
    M ./tests/two_dimensional.cpp -1 +1

Fri Apr 25 15:42:38 EDT 2008  [e-mail]
  * is-integrated?, not is-integrated

    M ./libctl/meep.scm.in -4 +4

Thu Apr 24 21:08:03 EDT 2008  [e-mail]
  * added store-polarization-energy? input variable to support that feature of the polarizations

    M ./libctl/meep.scm.in -3 +5

Thu Apr 24 21:01:49 EDT 2008  [e-mail]
  * disable slow, undocumented saturable-absorber support by default; use simple hand-written alternative in update_from_e.cpp; computingg polarization energy is now optional (to save memory and time since it is rarely used in practice)

    M ./configure.ac +7
    M ./src/fields.cpp -7 +8
    M ./src/meep.hpp -3 +3
    M ./src/meep_internals.hpp -3 +4
    M ./src/polarization.cpp -4 +11
    M ./src/update_e_from_d.cpp -11 +13
    M ./src/update_from_e.cpp -1 +34
    M ./tests/known_results.cpp -1 +5

Thu Apr 24 20:55:17 EDT 2008  [e-mail]
  * typo

    M ./configure.ac -1 +1

Thu Apr 24 20:54:49 EDT 2008  [e-mail]
  * fix thermal_energy to include all polarization contributions

    M ./src/energy_and_flux.cpp -4 +5

Thu Apr 24 20:54:26 EDT 2008  [e-mail]
  * fixed cw_solve to work with recent patch for non-integrated currents

    M ./src/cw_fields.cpp -1 +1
    M ./src/meep.hpp -2 +2
    M ./src/step.cpp -5 +6

Thu Apr 24 20:40:32 EDT 2008  [e-mail]
  * disable checks for floating-point exceptions in debugging mode, as some versions of g++ seem to have spurious exceptions in the complexx number routines(?)

    M ./src/mympi.cpp -1 +1

Sun Apr  6 20:28:31 EDT 2008  [e-mail]
  * support HDF5 1.8

    M ./src/h5file.cpp +4

Sun Apr  6 13:56:43 EDT 2008  [e-mail]
  * more robust link check for guile when guile-config not available

    M ./configure.ac -1 +2

Sun Apr  6 13:54:52 EDT 2008  [e-mail]
  * no way to test how to activate readline when cross-compiling

    M ./configure.ac +2

Sun Jan  6 16:37:49 EST 2008  [e-mail]
  * add to TODO

    M ./TODO +6

Fri Dec 21 20:05:17 EST 2007  [e-mail]
  * fix h5 filename timestamps for cases where dt is very small or very large (just use the integer time step number in that case)

    M ./src/h5fields.cpp -1 +6

Fri Dec 14 14:19:29 EST 2007  [e-mail]
  * check for ctl.h in /usr/include/ctl/, which is the default under Fedora (grrr)

    M ./configure.ac -1 +15
    M ./libctl/Makefile.am -1 +1

Thu Dec  6 20:03:02 EST 2007  [e-mail]
  * added is_integrated option to src_time to allow non-integrated E sources and non-integrated custom sources

    M ./libctl/meep.scm.in -18 +25
    M ./src/meep.hpp -2 +17
    M ./src/step.cpp -4 +2
    M ./src/update_e_from_d.cpp -2 +9

Tue Nov 13 15:27:02 EST 2007  [e-mail]
  tagged 0.10.1


Tue Nov 13 15:15:27 EST 2007  [e-mail]
  * release date for 0.10.1

    M ./NEWS -1 +1

Tue Nov 13 15:13:59 EST 2007  [e-mail]
  * update copyright year to 2007

    M ./COPYRIGHT -1 +1
    M ./libctl/meep.scm.in -1 +1
    M ./src/bands.cpp -1 +1
    M ./src/bicgstab.cpp -1 +1
    M ./src/bicgstab.hpp -1 +1
    M ./src/boundaries.cpp -1 +1
    M ./src/control_c.cpp -1 +1
    M ./src/cw_fields.cpp -1 +1
    M ./src/dft.cpp -1 +1
    M ./src/energy_and_flux.cpp -1 +1
    M ./src/fields.cpp -1 +1
    M ./src/grace.cpp -1 +1
    M ./src/h5fields.cpp -1 +1
    M ./src/h5file.cpp -1 +1